Schedule of FYP Proposal Evaluation (SP26)

1
Committee: 1
Team Members:
Dr. Kashif Ayyub (HEAD)
Ms. Yasmeen Khaliq
Mr. Ashfaq Ahmed
Ms. Maha Rasheed
Venue: CS Conference room
Remarks: Be on time

1{"id":1060,"project_id":1432,"title":"AR-Space Planner: A Real-Time Augmented Reality System for Interior Design and \r\nFurniture Visualization","prob":"Interior design and furniture purchasing largely depend on imagination, static images, \r\nand manual measurements, which often leads to poor decision-making. Customers are \r\nunable to accurately visualize how furniture will look, fit, or align with their existing \r\ninterior in terms of size, color, lighting, and layout. This visualization gap frequently \r\nresults in incorrect purchases, space miscalculations, and high return rates, causing \r\nfinancial loss for both customers and furniture sellers. The proposed system solves this \r\nproblem by providing a real-time, markerless Augmented Reality solution that allows \r\nusers to place and interact with virtual 3D furniture inside their actual environment using \r\na smartphone camera, ensuring accurate visualization before purchase.","description":"The AR-Space Planner application works by scanning the user\u2019s environment using a combination of the smartphone camera and gyroscope sensors to detect both horizontal and vertical planes, such as floors and walls, with high precision. Visual indicators guide the user through this initialization, leveraging inertial data to maintain stability even during rapid movement. Once surfaces are identified, the user selects furniture or wall-mounted items from categorized menus and uses raycasting to place 3D models accurately on the chosen plane. Users can manipulate these items via touch gestures while the system maintains realistic lighting and shadows. To document the design, the app includes a media capture suite that allows users to take high-resolution snapshots or record video walkthroughs of their virtual layout, which are saved directly to the device. Users retain full control to remove items or reset the scene at any time.$$\n1. Environment Scanning Module \r\n2. Furniture Catalog Module \r\n3. Object Placement & Interaction Module \r\n4. Lighting & Rendering Module\r\n5. Surface detection and mapping using Gyroscope sensor of devices$$\nI will develop the Environment Scanning and Plane Detection Module, which includes \r\ncamera initialization, real-time surface detection, plane visualization, and user guidance \r\nfor scanning indoor spaces using AR Foundation.$$\nI will develop the Furniture Catalog and Object Placement Module, which includes \r\nfurniture selection UI, raycasting-based object placement, and spawning optimized 3D \r\nmodels into the detected real-world environment.$$\n$$\n1. (Bajwa Furniture) Furniture Store with visual searching and recommendation system \r\n2. Augmented Reality Based Furniture Store Mobile Application \r\n3. RoomLoom - An AI-Powered Furniture Advisor$$\nAI model used in our FYP detects light and renders the object according to desired room and space.$$\nOur FYP is multi-platform Based i.e. it will work on Android as well as Apple devices$$\nOur FYP includes both AI module as well as Augmented reality concept for better Clearance and justification.","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-10 02:49:27","updated_at":"2026-03-24 15:55:49","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1061,"project_id":1433,"title":"HEMS \u2013 Hybrid Emergency Messaging System with Resilient Mesh Networking and Adaptive \r\nRelay Management","prob":"During natural disasters, infrastructure failures, and remote-area scenarios, cellular networks and \r\ninternet services become unavailable, leaving people unable to communicate, share critical \r\ninformation, or request emergency assistance. Existing messaging applications depend entirely \r\non centralized infrastructure such as cell towers and Wi-Fi routers, making them completely non\r\nfunctional when connectivity is lost. HEMS addresses this problem by enabling reliable \r\nencrypted communication even without any network infrastructure. The system creates a self\r\nhealing mesh network using Bluetooth Low Energy where every nearby device automatically \r\nrelays messages through multiple hops. Adaptive relay selection dynamically distributes \r\nforwarding responsibility based on device battery levels and connection stability. If any device \r\nregains internet access, it acts as a gateway to deliver messages through the backend server. \r\nStore-and-forward queuing, heartbeat-based failure detection, and alternate path routing ensure \r\nthe communication chain never breaks, even when intermediate devices disconnect.","description":"HEMS is an Android-based mobile communication platform that provides reliable encrypted \r\nmessaging during network outages and emergency situations using a hybrid approach combining \r\noffline mesh networking with opportunistic internet gateways. When no internet connectivity is \r\navailable, the application activates Bluetooth Low Energy discovery to detect nearby devices \r\nrunning HEMS. Discovered devices form a temporary mesh network where every device \r\nfunctions as both a sender and relay node. Messages are forwarded through multi-hop routing \r\nuntil reaching the destination or an internet-connected gateway device. The system implements \r\nadaptive relay management where each device shares battery level and connection stability \r\ninformation through piggybacked metadata in regular packets, requiring zero additional network \r\noverhead. Routing decisions prioritize devices with higher battery and stronger connections as \r\nrelay nodes. When battery drops below a configurable threshold, the device automatically \r\nreduces scanning frequency and declines non-emergency relay requests while still forwarding \r\ncritical emergency messages. For network resilience, HEMS implements heartbeat-based node \r\nfailure detection, alternate path routing through disjoint routes, and store-and-forward queuing. \r\nWhen a relay device disconnects, the system detects the failure within thirty seconds and \r\nautomatically reroutes queued messages through alternate peers. End-to-end encryption using \r\nhybrid RSA and AES cryptography ensures intermediate relay devices cannot access message \r\ncontent. The FastAPI backend manages message synchronization, delivery acknowledgments, \r\nand encrypted storage for offline recipients. An SOS emergency module enables panic alerts with \r\nlive GPS tracking and automated notifications through both mesh broadcast and SMS. A \r\nvalidation framework with automated failure scenario testing verifies system reliability under \r\nsimulated disaster conditions.$$\nThe Hybrid Emergency Messaging System with Resilient Mesh Networking and Adaptive Relay Management (HEMS) consists of the following modules: \r\n1. SecureGate (User Authentication & Identity Module) \r\nManages user registration, phone-based OTP verification, login\/logout, and JWT session management. Each device receives a unique cryptographic identity with RSA-4096 key pair generation for secure communication. The module manages credential storage, device-to-user \r\nmapping, and authentication token lifecycle. \r\n2. CipherChat (Encrypted Messaging Module) \r\nHandles message creation, hybrid RSA\u2013AES end-to-end encryption, transmission, and delivery tracking across online and offline channels. Messages are encrypted on the sender device and can \r\nonly be decrypted by the intended receiver. Relay nodes only forward encrypted packets. The module maintains queued, sent, delivered, and failed states and verifies integrity using SHA-256 hashing. \r\n3. MeshLink (Mesh Communication & Discovery Module) \r\nEnables device-to-device communication using Bluetooth Low Energy through the Google Nearby Connections API when internet connectivity is unavailable. It manages device discovery, peer connections, data transfer, and reconnection, allowing nearby devices to form a \r\ndecentralized mesh network where each node acts as a relay. \r\n4. PathFinder (Adaptive Routing & Relay Management Module) \r\nManages multi-hop message forwarding using routing strategies such as flooding, gossip, and adaptive routing. Relay selection considers battery level and connection stability. The module controls hop count, TTL, unique message IDs, loop prevention, priority handling, and alternate \r\npath routing. \r\n5. ShieldNet (Failure Recovery & Network Resilience Module) \r\nEnsures communication reliability through heartbeat-based node monitoring and automatic rerouting when nodes disconnect. It also implements store-and-forward messaging, where undelivered messages are stored locally and retransmitted when new routes become available. \r\n6. PowerGuard (Energy Optimization Module) \r\nImplements battery-aware networking behavior. Devices with low battery reduce scanning frequency and limit relay duties for non-critical messages, while emergency messages are always forwarded. Relay load is distributed toward higher-battery devices to improve network longevity. \r\n7. CloudBridge (Gateway & Backend Services Module) \r\nDetects devices with internet connectivity that act as gateway nodes. These nodes forward encrypted messages to the FastAPI backend server, which manages authentication APIs, message synchronization, offline message queues, and WebSocket updates. \r\n8. PanicBeacon (SOS Emergency & GPS Module) \r\nProvides a panic button for emergency alerts. When triggered, it sends GPS coordinates to the backend and broadcasts alerts through the mesh network while notifying emergency contacts. \r\n9. ProofEngine (Validation & Testing Framework Module) \r\nValidates system reliability through failure scenario simulations such as node failure, network partition, multi-hop routing, and TTL expiry. Real-device tests measure delivery success rate, latency, and rerouting performance. \r\n10. DataVault (Database & Persistence Module) \r\nManages local and backend storage. Mobile devices use Hive and Secure Storage for queued messages and encryption keys, while the backend uses PostgreSQL for encrypted messages, authentication records, and SOS data$$\nIn HEMS I shall develop the following modules: (i) MeshLink (Mesh Communication & Discovery \r\nModule) (ii) PathFinder (Adaptive Routing & Relay Management Module) (iii) CipherChat (Encrypted \r\nMessaging Module) (iv) DataVault (Database & Persistence Module) (v) Mobile Application Interface. \r\nThis includes implementing Bluetooth Low Energy communication through Google Nearby Connections \r\nAPI, peer discovery and connection management, multi-hop message forwarding with adaptive relay \r\nselection based on battery and stability scoring, loop prevention using message IDs and TTL, and mobile\r\nside encryption integration. I will also develop the PanicBeacon mobile interface including the SOS panic \r\nbutton, GPS tracking screen, and emergency type selection, along with store-and-forward local message \r\npersistence.$$\nIn HEMS I shall develop the following modules: (i) SecureGate (User Authentication & Identity Module) \r\n(ii) CloudBridge (Gateway & Backend Services Module) (iii) PowerGuard (Energy Optimization \r\nModule) (iv) ShieldNet (Failure Recovery & Network Resilience Module) (v) ProofEngine (Validation & \r\nTesting Framework Module). This includes implementing FastAPI backend services with authentication \r\nAPIs, JWT token management, encrypted message storage, gateway detection logic, WebSocket real-time \r\ncommunication, heartbeat-based node failure detection, automatic rerouting through alternate paths, \r\nbattery-aware relay scoring, and the automated failure scenario simulation framework. I will also \r\nconfigure Docker-based deployment with CI pipelines and develop the backend SOS emergency service \r\nAPIs.$$\n$$\n$$\nBattery-Optimized Adaptive Relay Selection \u2014 Devices select relay nodes using \r\nBATTERY LEVEL and connection stability across the MESH network without master\r\nslave centerlized i.e server dependant$$\nAutomatic Failure Recovery with Alternate Path Routing \u2014 Heartbeats detect node \r\nfailures, enabling alternate routing and message preservation during partitions and \r\nfallback mechanism$$\nReliability Validation Framework \u2014 System implementation done by real-world testing \r\nfailure scenarios such as node failure, network partition, TTL expiry, and priority \r\nmessage handling","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-10 14:18:10","updated_at":"2026-03-11 20:08:52","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1065,"project_id":1434,"title":"SafeToon: AI-Powered Parental Control and Smart Cartoon Recommendation System for Children on YouTube.","prob":"Children increasingly consume video content on YouTube without proper supervision, where the platform\u2019s recommendation algorithms are mainly optimized for engagement rather than child safety or educational value. As a result, children may be exposed to age-inappropriate, misleading, violent, or addictive content. Although basic parental control features exist, they often fail to accurately filter videos due to clickbait titles, misleading tags, and algorithm-driven recommendations. Parents therefore face difficulty in monitoring what their children watch and ensuring that the content is safe and beneficial for learning. This project aims to address this problem by providing an AI-based system that can analyze video metadata, filter inappropriate content, and recommend age-appropriate cartoons and educational videos. The proposed solution will help parents guide children toward safer and more meaningful digital media consumption.","description":"The proposed Final Year Project, SafeToon, is a web-based system designed to provide children with safe, age-appropriate cartoon and educational video recommendations while giving parents better control over the content their children watch online. With the increasing use of digital platforms such as YouTube, children often encounter content that may not be suitable for their age group. The main objective of this system is to create a safe and intelligent recommendation environment that filters and suggests appropriate videos for children.\r\nThe system will allow parents to create and manage a child profile by specifying basic information such as age group and content preferences. The platform will maintain a curated database of cartoon and educational videos collected from YouTube through data gathering and filtering techniques. Using machine learning\u2013based recommendation methods, the system will analyze video metadata such as titles, descriptions, categories, and engagement indicators to recommend suitable cartoons and learning videos.\r\nTo make the experience engaging for children, the system will provide a simple and fun interface where children can select their mood (e.g., happy, sad, excited) or preferred cartoon category. Based on this input, the system will generate personalized recommendations. When a child selects a video, the system will redirect the user to the safe version of the video on YouTube for playback.\r\nAdditionally, the system will include a parental control feature through a browser extension that allows parents to monitor and guide children\u2019s viewing behavior. The overall system will consist of modules for user management, content filtering, recommendation generation, parental monitoring, and data management, ensuring a secure and educational entertainment experience for children.$$\n1. User Management Module\r\nThis module handles the registration, login, and profile management of users. Parents can create accounts and add child profiles by specifying information such as age group and preferences. The module also manages authentication and authorization to ensure that only authorized users can access the system. It maintains user records and stores relevant information in the database for personalization.\r\n\r\n2. Child Interaction Module\r\nThis module provides a simple, colorful, and child-friendly interface where children can interact with the system. Instead of complex searching, children can select their mood (e.g., happy, excited, relaxed) or choose cartoon categories. The system then processes the selected option and forwards it to the recommendation engine to generate appropriate suggestions.\r\n\r\n3. Video Data Collection Module\r\nThis module is responsible for collecting cartoon and educational video information from YouTube. The system gathers video metadata such as titles, descriptions, categories, duration, and popularity indicators. This data is then stored in the system\u2019s database for further filtering and recommendation processing.\r\n\r\n4. Content Filtering Module\r\nThe purpose of this module is to ensure that only safe and age-appropriate content is available to children. The module analyzes video metadata and removes videos that contain inappropriate keywords or categories. It acts as a protective layer before videos are recommended to users, improving the safety of children\u2019s digital viewing experience.\r\n\r\n5. Recommendation Engine Module\r\nThis is the core intelligent component of the system. It uses machine learning or rule-based recommendation techniques to analyze children\u2019s preferences, mood selections, and video metadata. Based on these factors, the system generates personalized cartoon and educational video recommendations that match the child\u2019s interests and age group.\r\n\r\n6. Video Redirection Module\r\nOnce a recommendation is selected, this module redirects the user to the safe playback page of the video on YouTube. Instead of hosting the videos directly, the system provides controlled access by linking children to filtered and verified content.\r\n\r\n7. Parental Control & Monitoring Module\r\nThis module allows parents to monitor the videos recommended to their children. Parents can view the watch history, manage child preferences, and guide content consumption. The module may also include a browser extension to help parents control what children watch while browsing.\r\n\r\n8. Database Management Module\r\nThis module manages the storage and organization of all system data, including user accounts, child profiles, video metadata, and recommendation logs. It ensures efficient data retrieval and secure storage for the smooth functioning of the entire system.$$\nI shall develop the Recommendation Engine, Content Filtering, Video Data Collection, and Database Management modules. These modules will be responsible for collecting cartoon and educational video data from YouTube, storing the data in the system database, filtering inappropriate content, and generating personalized recommendations for children based on their age group, preferences, and selected mood. This part represents the core intelligent functionality of the system.$$\nThe second member shall develop the User Management, Child Interaction Interface, Video Redirection, and Parental Control & Monitoring modules. These modules will manage user registration and login, provide a child-friendly interface for selecting moods or categories, redirect recommended videos to YouTube for playback, and allow parents to monitor and manage their child\u2019s viewing preferences and activity.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-11 17:22:56","updated_at":"2026-03-11 17:22:56","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1085,"project_id":1441,"title":"Quit-Companion: A mobile app supporting nicotine recovery through habit tracking, personalized guidance, improved overall health.","prob":"Nicotine addiction is increasingly common among young adults and university students through cigarettes, vapes, and nicotine pouches. Many individuals want to quit struggle due to psychological stress, social triggers, cravings, and lack immediate support. Existing quitting applications often generic and non-personalized, failing to combine behavioral analysis with health monitoring. As result, users frequently relapse because they cannot understand triggers or receive timely assistance.\r\n\r\nQuit-Companion addresses these gaps by providing an AI-powered, data-driven, medically informed support system. To ensure data legitimacy, the application will use verified public datasets: NHANES Tobacco Use Dataset: https:\/\/www.cdc.gov\/nchs\/nhanes\/\r\n , SAMHSA National Survey on Drug Use and Health (NSDUH): https:\/\/www.samhsa.gov\/data\/\r\n , and UCI Machine Learning Repository \u2014 Smoking and Drinking Dataset: https:\/\/archive.ics.uci.edu\/dataset\/609\/\r\n . These datasets include validated data on smoking behavior, demographics, and health indicators. Additionally, voluntary user data (daily usage, cravings, mood patterns, withdrawal symptoms) will help improve model accuracy.","description":"Quit-Companion is a mobile application designed to support individuals in overcoming nicotine addiction through a structured, research-driven, and technology-supported approach. The application assists users throughout their quitting journey by combining habit tracking, trigger monitoring, health awareness, and motivational support in a single platform. Instead of forcing abrupt quitting, the system encourages gradual reduction of nicotine consumption while helping users understand the behavioral, psychological, and environmental factors that influence their addiction.\r\nThis approach strengthens the project\u2019s scope by integrating behavioral analysis, AI-driven insights, and continuous user engagement throughout the quitting process.\r\n\r\nTo address the committee\u2019s concern regarding research-based suggestions, the system will use a fine-tuned AI model trained on pre-existing validated datasets related to nicotine addiction and health behavior. Training the model on reliable datasets ensures that the recommendations provided to users are grounded in real-world research and evidence rather than generic suggestions.\r\n\r\nTo develop personalized quitting strategies, the system will also implement a structured mechanism for collecting user data. With the user\u2019s consent, the application will record information such as daily nicotine usage, craving frequency, common triggers, mood changes, and withdrawal symptoms. This data will allow the system to identify addiction patterns and continuously refine the AI model to generate personalized recommendations for each user.\r\n\r\nAnother major concern addressed is medical supervision and safety. The project will be developed under the guidance of Dr. Bushra Ali(MBBS , RMP ,MD ,Trainee) who will act as the domain expert. She will review the health-related guidance generated by the AI and the Health-Bot module to ensure that all recommendations comply with professional medical standards and avoid any potentially harmful advice.\r\n\r\nBy integrating AI-based personalization, structured user data collection, and professional medical guidance, Quit-Companion expands its scope beyond a simple tracking application and becomes a safe and research-supported digital support system for nicotine cessation.$$\nThe system contains total nine modules.\r\n1. User Onboarding & Profiling Module\r\nThis module is responsible for collecting essential information when a user first registers on the application. The user provides details such as the type of nicotine product used (vape, cigarettes, pouches), daily consumption, motivation to quit, and personal challenges. All collected data is securely stored in Firebase, ensuring privacy and data protection. Based on this information, the system sets personalized quitting goals, creating a tailored experience for each user from the beginning.\r\n\r\n2. Dashboard & Progress Analytics Module\r\nThe dashboard serves as the main interface where users can track their quitting progress. It displays real-time statistics such as nicotine-free days, money saved, and the number of cravings avoided. Visual charts and graphs help users easily understand their progress. Additionally, a quitting timeline is generated using milestones, which motivates users by showing how far they have come and what goals are ahead.\r\n\r\n3. Habit & Trigger Tracking Module\r\nThis module allows users to log their cravings along with the reasons, mood, time, and activities during the craving. The system records the frequency and timing of cravings. Using AI-based analysis, the application identifies common triggers such as stress, social situations, or boredom. It also predicts relapse risks by recognizing recurring patterns, enabling early intervention.\r\n\r\n4. Alternatives & Habit Replacement Module\r\nTo help users manage cravings instantly, this module provides healthy alternatives such as chewing gum, breathing exercises, drinking water, or engaging in mini-games. The AI system analyzes which alternatives work best for the user and recommends the most effective options based on past success, making craving management more efficient over time.\r\n\r\n5. Health-Bot (Symptom Checker)\r\nThe Health-Bot is a conversational chatbot that answers user health-related questions. It explains common withdrawal symptoms such as headaches, irritability, or sleep issues and provides safe guidance. If a user reports serious or concerning symptoms, the bot recommends consulting a medical professional instead of providing risky advice.\r\n\r\n6. Gamification & Rewards Module\r\nThis module keeps users motivated by introducing points, streaks, and badges. Users earn rewards for achieving milestones such as 24 hours, 7 days, or 1 month without nicotine. Gamification increases engagement and makes the quitting journey more enjoyable and goal-oriented.\r\n\r\n7. Community & Peer Support Module\r\nThe application includes a community feature where users can interact through forums and chat. Users share experiences, success stories, and tips, creating a supportive environment. Peer encouragement plays a crucial role in reducing relapse chances and increasing motivation.\r\n\r\n8. Relapse Prevention & Reflection Module\r\nUsers can maintain journals to reflect on their quitting journey. In case of relapse, the AI analyzes the triggers and provides a personalized restart plan. The system also generates summaries and practical suggestions to help prevent future relapses.\r\n\r\n9. Knowledge Hub & Education Module\r\nThis module provides educational resources such as articles, videos, and health facts related to nicotine addiction. It educates users about the harmful effects of nicotine and the long-term health benefits of quitting, empowering them with knowledge to stay committed.$$\n1. User Onboarding & Profiling Module\r\no Collects initial information about the user such as nicotine type, daily \r\nusage, motivation, and challenges.\r\no Stores user data securely in Firebase.\r\no Sets personalized goals.\r\n2. Dashboard & Progress Analytics Module\r\no Displays real-time progress including nicotine-free days, money saved, \r\nand cravings avoided.\r\no Shows visual charts and statistics.\r\no Generates a quitting timeline based on milestones.\r\n3. Habit & Trigger Tracking Module\r\no Users log cravings, reasons, mood, and activities.\r\no System records time and frequency of cravings.\r\no AI analyzes patterns to identify common triggers and predict relapse risk.\r\n4. Alternatives & Habit Replacement Module\r\no Provides healthy alternatives such as chewing gum, breathing exercises, \r\nwater intake, or mini-games.\r\no Helps users manage cravings instantly.\r\no AI recommends and the most effective alternatives based on past user \r\nsuccess.\r\n9. Knowledge Hub & Education Module\r\no Includes articles, videos, and facts.$$\n4. Alternatives & Habit Replacement Module\r\no Provides healthy alternatives such as chewing gum, breathing exercises, \r\nwater intake, or mini-games.\r\no Helps users manage cravings instantly.\r\no AI recommends the most effective alternatives based on past user \r\nsuccess.\r\n5. Health-Bot (Symptom Checker)\r\no A conversational chatbot answers user health queries.\r\no Explains withdrawal symptoms provides safe guidance.\r\no Recommends visiting a doctor for serious conditions.\r\n6. Gamification Rewards Module\r\no Motivates users using points, streaks, and badges.\r\no Users earn rewards for achieving milestones (24 hours, 7 days, 1 month).\r\n7. Community & Peer Support Module\r\no Users can communicate with others through forums and chat.\r\no They share experiences, tips, and support each other.\r\n8. Relapse Prevention & Reflection Module\r\no Users write journals about their quitting journey.\r\no If relapse occurs, AI analyzes triggers provides a personalized restart \r\nplan.\r\no Generates summaries and suggestions to prevent future relapse.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-16 17:00:04","updated_at":"2026-03-24 15:56:11","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1064,"project_id":1442,"title":"AI-Assisted Augmented Reality System for Product Size and Fit Visualization in Online\r\nShopping","prob":"Online shopping is convenient but customers often cannot clearly understand the actual size and fit of products before buying them. Products are usually shown only through images, so buyers cannot judge whether items like furniture or appliances will properly fit in their room or space. As a result, customers sometimes purchase products that are too large, too small, or unsuitable, which leads to product returns, extra delivery costs, and customer dissatisfaction. This FYP solves this problem by providing a system that helps users visualize products in their real environment. The main aim is to find the best and correct size and fit before making a purchase decision.","description":"This project develops a mobile Augmented Reality (AR) application that helps users visualize products in their real environment before purchasing them online. The system allows users to scan their room using the mobile camera and then place virtual 3D models of products in true-to-scale size within that space. An AI-assisted recommendation system analyzes the detected space and suggests the most suitable product size based on available area and user preferences. By allowing users to see how products like furniture, appliances, or d\u00e9cor will look and fit in their environment, the system helps customers make better purchasing decisions and reduces the chances of buying wrong-sized products or returning them.$$\n1. Room Scanning Module\r\nThis module uses the mobile device camera to scan the user\u2019s room and detect the available space where a product can be placed.\r\n\r\n2. Product Catalog Module\r\nThis module displays a list of available products with their 3D models and details so the user can select a product to visualize.\r\n\r\n3. AR Visualization Module\r\nThis module places the selected product as a true-to-scale 3D model in the user\u2019s real environment using Augmented Reality, allowing the user to view it from different angles.\r\n\r\n4. AI Recommendation Module\r\nThis module analyzes the room dimensions and available space and suggests the most suitable product size for the user.\r\n\r\n5. User Interaction Module\r\nThis module allows users to move, rotate, and adjust the virtual product in the room to better understand its placement and fit.$$\nI will develop the Room Scanning Module and AR Visualization Module. These modules involve implementing the AR system, integrating the mobile camera for room scanning, and placing true-to-scale 3D product models in the user\u2019s real environment. This work mainly focuses on system development and AR functionality, which aligns with her software engineering background.$$\nI will develop the Product Catalog Module, AI Recommendation Module, and User Interaction Module. These modules include managing product data, storing and retrieving product information from the database, and implementing AI-based recommendations based on room size and user preferences.$$\n$$\nVISION: A QR Code Scanning Application for Augmented Reality Object Visualization \r\nAR Designing \u2013 Application for Augmented Reality Designing$$\nAI-Based Product Size Recommendation\r\nThe system analyzes the available room space and suggests the most suitable product size so that it fits properly in the user\u2019s environment.$$\nTrue-to-Scale AR Product Visualization\r\nUsers can place 3D models of products in their real environment using Augmented Reality, allowing them to see the product in actual size before purchasing.$$\nInteractive Product Placement\r\nUsers can move, rotate, and adjust the virtual product in their room to check different positions and understand how well it fits in the available space.","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-11 16:10:39","updated_at":"2026-03-18 09:34:08","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1084,"project_id":1446,"title":"Skill-Nest \u2013 An AI-powered platform that grows your skills and guides to perfect job","prob":"Many fresh graduates struggle to find suitable jobs due to unclear skill gaps, ineffective resumes, and lack of guidance on industry requirements. They often don\u2019t know which projects to build, courses to take, or certifications to pursue. Skill Nest solves these problems by providing an AI-powered career assistant that analyzes resumes,suggest his skill related jobs, detects missing skills, recommends personalized projects and learning paths, predicts success probability for jobs, estimates expected salaries, and optimizes resumes to improve employability and match industry needs..","description":"The proposed system, Skill nest, is an AI-powered platform designed to help fresh graduates and entry-level job seekers enhance their employability. Users upload their resumes, which are analyzed using AI to extract skills, education, experience, and projects. The system identifies skill gaps, recommends personalized learning paths, and suggests real-world projects to build practical experience. It matches users to suitable jobs, calculates the probability of success for each role, predicts expected salaries, and optimizes resumes to improve shortlisting chances. A career dashboard consolidates insights, including skill evolution, job market trends, and competitive comparisons. Overall, the system acts as an intelligent career mentor, guiding students from skill assessment to job readiness.$$\nIn Skill Nest, the modules are:\r\n(i) User Registration & Profile Management \u2013 Handles secure registration, authentication, and profile creation for users\r\n(ii) Resume Upload & Parsing \u2013 Extracts structured information from resumes including skills, education, experience, and projects using AI.\r\n(iii) Job Data Collection & Skill Database \u2013 Gathers job postings from APIs and builds a database of required skills and roles.\r\n(iv) AI Job Matching Engine \u2013 Matches user skills with job requirements and calculates job fit scores.\r\n(v) Skill Gap Analysis & Recommendation \u2013 Identifies missing skills and suggests personalized learning paths.\r\n(vi) AI Resume Optimization \u2013 Tailors resumes for specific job applications to improve chances of shortlisting.\r\n(vii) Personalized Project Generator \u2013 Recommends real-world projects with problem statements, tools, datasets, and difficulty levels to build practical experience.\r\n(viii) Job Success Probability & Salary Predictor \u2013 Predicts likelihood of selection for specific jobs and estimates expected salary based on profile and market data.\r\n(ix) Career Dashboard & Analytics \u2013 Consolidates insights including skill evolution, job market trends, competitive resume comparison, recommended courses, projects, and application tracking.\r\n(x)Interview Guidance Module-To provide real-time guidance, interview preparation, and career advice to users.$$\nIn Skill Nest, I shall develop these modules:vi) AI Resume Optimization(vii) Personalized Project Generator(viii) Job Success Probability & Salary Predictor(ix) Career Dashboard & Analytics(x)Skill Improvement Recommendation$$\nIn Skill Nest, I shall develop these modules: (i) User Registration & Profile Management(ii) Resume Upload & Parsing(iii) Job Data Collection & Skill Database(iv) AI Job Matching Engine(v) Skill Gap Analysis(vi)Interview Guidance Module$$\n$$\nCareer Bridge\r\nCareer Assist$$\n1. AI-Powered Career Guidance$$\n2. Personalized Learning & Project Recommendations$$\n3. Salary Prediction & Career Dashboard","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-15 22:40:57","updated_at":"2026-03-15 22:40:57","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1056,"project_id":1447,"title":"Web-Based Enterprise Resource Planning (ERP) System for Young\u2019s Food Pvt Ltd","prob":"Young\u2019s Food Pvt Ltd is a manufacturing organization involved in the production and distribution of food products such as biscuits, wafers, and confectionery items. The company currently manages its operational activities using manual records and separate digital tools across different departments.\r\n\r\nThis traditional approach has resulted in several operational challenges, including:\r\n\r\nLack of centralized data management across departments\r\n\r\nInventory inaccuracies and difficulty in tracking raw materials and finished products\r\n\r\nDelays in accessing employee and sales information\r\n\r\nTime-consuming manual report preparation\r\n\r\nPoor coordination between inventory, sales, and administrative departments\r\n\r\nLimited real-time visibility of organizational performance\r\n\r\nDue to these issues, management faces difficulty in monitoring daily operations and making informed decisions efficiently. These problems reduce productivity and may lead to stock shortages, overstocking, and operational inefficiencies.\r\n\r\nThe proposed project aims to address these challenges by developing an integrated digital management system tailored to the operational needs of Young\u2019s Food Pvt Ltd.","description":"The proposed project focuses on developing a web-based Enterprise Resource Planning (ERP) system for Young\u2019s Food Pvt Ltd to integrate and manage its core business operations within a centralized platform.\r\n\r\nThe system will digitize organizational processes including employee management, inventory tracking, sales management, and performance reporting. Authorized users from different departments will access the system through role-based dashboards designed according to their responsibilities.$$\n1. Authentication & Role Management\r\n\r\nThis module provides secure login functionality and manages user access based on assigned roles. It ensures that users can only access system features related to their responsibilities.\r\n\r\n2. Employee Management\r\n\r\nThis module maintains employee information including personal details, department records, and attendance data, enabling centralized workforce management.\r\n\r\n3. Inventory & Stock Management\r\n\r\nThis module manages raw materials and finished product stock. It allows monitoring of inventory levels, updating stock records, and maintaining accurate inventory information.\r\n\r\n4. Sales Management\r\n\r\nThis module records sales transactions, manages distributor or customer information, and maintains sales and payment records.\r\n\r\n5. Analytics & Reporting Dashboard\r\n\r\nThis module provides summarized reports and graphical dashboards to monitor organizational performance such as sales trends and inventory status.$$\nAuthentication & Role Management Module\r\n\r\nEmployee Management Module\r\n\r\nDatabase Design and Integration\r\n\r\nBackend API development for assigned modules\r\n\r\nRole:\r\nResponsible for implementing secure system access and managing employee-related operations along with database structure development.$$\nInventory & Stock Management Module\r\n\r\nSales Management Module\r\n\r\nAnalytics & Reporting Dashboard\r\n\r\nRole:\r\nResponsible for handling operational modules related to inventory tracking, sales processing, and dashboard visualization$$\n$$\n$$\n1. Centralized Integrated Management System\r\n\r\nThe system integrates multiple organizational functions such as employee management, inventory control, and sales management into a single centralized plat$$\n2. Role-Based Secure Access Control\r\n\r\nA role-based authentication mechanism will be implemented to provide secure system access. Different users (such as administrators, HR staff, and inventory manager$$\n3. Real-Time Analytics and Reporting Dashboard\r\n\r\nThe system will provide interactive dashboards and automated reports that present organizational data through charts and summaries. This functionality e","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-04 23:24:00","updated_at":"2026-03-04 23:24:00","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1081,"project_id":1485,"title":"CampusShed AI \u2013 Next\u2011Gen Scheduling (Industrial Project)","prob":"Universities struggle with manual timetable creation, causing classroom clashes, teacher and student overload, resource conflicts, and limited flexibility. Existing systems in different universities often lack AI-driven optimization, historical data analysis, pre-semester slot selection, and collaborative workflows. This results in wasted resources, scheduling conflicts, last-minute adjustments, and poor coordination between students, teachers, and CRs. Many universities also follow different academic workflows, role structures, and rescheduling policies, therefore the platform is designed to be flexible and supports configurable scheduling rules, customizable role structures, and adaptable coordination workflows so that each university can define its own scheduling process according to its institutional practices. Our system addresses these issues through AI-based timetable generation, pre-semester slot locking, teacher availability management, real-time conflict prevention, extra class and CR-request handling, and seamless communication via notifications and chatrooms. The system ensures secure authentication, role-based access, multi-university data isolation, and scalable performance to handle multiple users simultaneously.","description":"This system is a comprehensive SaaS-based university timetable management and\r\ncoordination platform that automates schedule generation using AI algorithms\r\ntrained on historical data. It supports multi-tenant architecture for university\r\nisolation and role-based access for coordinators, teachers, and CRs. The AI engine\r\nanalyzes past semesters to recommend 2\u20133 optimized timetables, minimizing\r\nconflicts while balancing teacher and student workloads. Teachers select slots\r\nthrough a pre-semester locking system on a first-come, first-serve basis and mark\r\nvisiting slots for student consultation bookings. The platform supports configurable\r\nacademic workflows and scheduling policies so different universities can define\r\ntheir own coordination structures and class rescheduling processes according to\r\ntheir institutional practices. CRs and teachers submit timetable change requests via\r\ncoordinator approval workflows. The system also manages extra class\r\narrangements, student appointment bookings, CR-requested teacher bookings, and\r\nchat-based communication, with automatic conflict resolution, room allocation,\r\nand capacity-aware scheduling. The system also ensures secure authentication,\r\nrole-based access control, university-level data isolation, and scalable architecture\r\nto support large numbers of users while maintaining system performance.$$\nModule 1: Core Scheduling Engine & AI Intelligence System\r\nAutomatic timetable generation with conflict resolution\r\nIntelligent timetable recommendations with ranking algorithms\r\nTeacher and student workload balancing and room utilization optimization\r\nConfigurable scheduling rules for different university policies\r\n\r\nModule 2: User Management & Access Control System\r\nMulti-role authentication (Coordinator, Teachers, CRs)\r\nRole-based permissions and approval workflows\r\nMulti-tenant architecture for university isolation\r\nCR governance system with voting and polling features\r\nSecure authentication and university-level data isolation\r\n\r\nModule 3: Academic Configuration & Setup Module\r\nPre-semester slot locking mechanism with FCFS allocation\r\nManual coordinator override capabilities\r\nUniversity working hours and break management\r\nFlexible class duration support (45 min to 3 hours)\r\nSemester-based data management with historical archiving\r\nClassroom type handling (labs vs theory rooms)\r\nCredit hour and section management\r\nConfigurable academic workflows for different universities\r\n\r\nModule 4: Teacher Availability & Request Management\r\nFree slot availability marking system\r\nTeacher can update availability status if unavailable during free slots\r\nTeacher change request workflows (class movement, room change)\r\nAvailability constraint management\r\nCoordinator approval system with AI re-adjustment\r\nStudent appointment booking with teachers based on available visiting slots\r\n\r\nModule 5: CR Request & Extra Class Coordination\r\nCR timetable change requests with guided suggestions\r\nExtra class arrangement system with teacher\/CR collaboration\r\nReal-time room availability checking\r\nFCFS room allocation for simultaneous requests\r\nAutomatic alternate room and time slot suggestions\r\nFlexible rescheduling support for different university policies\r\n\r\nModule 6: Notification & Communication System\r\nChange request alerts and approval notifications\r\nReal-time workflow status updates\r\nChatroom for student teacher communication\r\n\r\nModule 7: Export & Integration Module\r\nPDF timetable generation (coordinator, teacher, class-specific)\r\nGoogle Calendar and Outlook export functionality\r\nExternal system integration\r\n\r\nModule 8: Analytics & Reporting Dashboard\r\nPerformance dashboard\r\nTeacher load analytics and distribution reports\r\nRoom utilization statistics and peak hour analysis\r\nConflict and change request trend tracking\r\nAI performance improvement metrics\r\nHistorical comparison reports\r\n\r\nModule 9: SaaS Platform & Commercial Features\r\nUniversity onboarding workflow\r\nCustom branding per university\r\nSubscription plan management\r\nUsage-based access limits\r\nFeature-based tier system\r\nScalable architecture for multi-university deployments$$\nIn CampusShed AI I shall develop Module 1: Core Scheduling Engine & AI Intelligence System, Module 3: Academic Configuration & Setup Module, Module 5: CR Request & Extra Class Coordination, Module 7: Export & Integration Module, and Module 9: SaaS Platform & Commercial Features.$$\nIn CampusShed I shall develop Module 2: User Management & Access Control System, Module 4: Teacher Availability & Request Management, Module 6: Notification & Communication System, and Module 8: Analytics & Reporting Dashboard.$$\n$$\nTime Table Generator$$\n1. AI-Powered Timetable Recommendation Engine$$\n2. Student appointment & CR-request management$$\n3. Real-time resource and communication system","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-15 15:13:26","updated_at":"2026-03-15 15:13:26","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1092,"project_id":1471,"title":"BOATMART \u2013 The All-in-One Solution for E-Commerce Module A","prob":"The e-commerce industry is highly fragmented: wholesalers, retailers, and dropshippers operate on separate platforms and rely on multiple disconnected tools for inventory, orders, customer management, payments, analytics, CRM, and ERP. This lack of integration increases operational costs, creates data inconsistency, slows workflows, and adds technical barriers\u2014especially for small and medium businesses\u2014making it harder to scale.\r\nThis FYP proposes a lightweight all-in-one e-commerce prototype that unifies key operations in a single system by connecting wholesalers, retailers, and dropshippers. The platform will centralize a shared product catalog, basic inventory visibility, order processing and routing, and simple status updates, along with supportive built-in tools such as basic CRM, sales analytics dashboards, reporting, and automation features. By reducing dependence on multiple tools, the solution aims to simplify workflows, improve efficiency, provide clearer data visibility, and lower operational complexity for growing businesses","description":"The proposed system is a unified e-commerce ecosystem designed to connect wholesalers, dropshippers\/sellers, and retail customers within a single integrated platform, eliminating the need for multiple tools and platforms.\r\nWholesalers register on the platform and access a dedicated supplier dashboard where they upload and manage product information, including pricing, variants, available stock, and estimated fulfillment times. Wholesalers do not sell directly to end customers; instead, they act as product suppliers for sellers on the platform.\r\nDropshippers or sellers create their own branded storefronts within the system. They can browse the unified product catalog, select products from registered wholesalers, and list them in their stores with customized pricing, branding, and product descriptions.\r\nRetail customers browse seller storefronts or a centralized marketplace, place orders, and track deliveries. Once an order is placed, the platform automatically routes it to the relevant wholesaler for fulfillment, manages order tracking, and handles secure payment distribution among the wholesaler, seller, and platform.\r\nThe system implements role-based authentication and authorization (customer, seller, wholesaler, administrator), ensuring secure access and controlled functionality for each user type.$$\nIn BoatMart, we are making Customer Module purpose is to provides customers with a secure, reliable, and user-friendly shopping experience.\r\nSub-Modules \r\n1. Conversational AI Commerce Assistant A smart chatbot that helps customers shop like a real person. Users can simply type or speak what they want and the AI understands the meaning to show the exact matching products.\r\n\uf0b7Natural Language Input Interface: A chat screen where users can type or send voice messages.\r\n\uf0b7Intent Extraction Engine: The AI that picks out important details like color, price, and category from the user's message.\r\n\uf0b7Semantic Search & Matcher: Searches the database using the meaning of the words, not just exact spellings.\r\n2. Automata-Driven Negotiation Engine A smart bargaining bot that lets customers negotiate prices just like in a physical market. \r\n\uf0b7Negotiation Chat Interface: A chat screen where customers can type the price they want to pay.\r\n\uf0b7State Machine Logic Controller: Manages the bargaining steps, from receiving an offer to giving a counter-offer.\r\n\uf0b7Wholesale Margin Calculator: A hidden checker that makes sure the bot never sells anything below the profit limit.\r\n3. Reinforcement Learning Discovery Pipeline A fun, Tinder-style shopping screen where customers swipe right to like an item or left to skip it. The AI learns their taste from these swipes and automatically starts showing products they will love.\r\n\uf0b7Swiping Card UI: A screen where users can easily swipe product cards left or right.\r\n\uf0b7User Action Logger: Records which items the user liked and which ones they ignored.\r\n\uf0b7RL Weight Updater: The AI brain that updates the user's style profile based on their swiping history.\r\n4. Multiplayer Social Commerce Gateway A feature that lets friends shop together at the same time. Users can share a cart link on WhatsApp, add items together, split the bill, and get special bulk discounts from wholesalers for ordering together.\r\n\uf0b7WebSocket Connection Manager: Keeps the cart updated live for everyone, without needing to refresh the page.\r\n\uf0b7Order Pooling Engine: Combines everyone's separate items into one big wholesale order.\r\n\uf0b7Wholesale Discount Calculator: Automatically lowers the price when the group orders enough items together.\r\n\uf0b7Bill Split & Checkout Calculator: Divides the total cost so each friend knows exactly what they have to pay.\r\n5. Algorithmic Trust & Sentiment Framework A system designed to build trust by showing honest reviews and reliable sellers. \r\n\uf0b7Sentiment Classification Model: An API that figures out if a review is positive, negative, or neutral.\r\n\uf0b7Seller Metrics Aggregator: Tracks how many orders a seller shipped on time and how many were returned.\r\n\uf0b7Real-time Score Updater: Automatically calculates and updates the seller's trust score on their profile.\r\n6. Predictive Analytics & Forecasting Hub A smart tracking system that guesses when a product's price will go down. It looks at past price changes and sends an alert to the customer before a discount happens, so they can save money.\r\n\uf0b7Watchlist Management Interface: A screen where users can save items they want to \r\n\uf0b7Time-Series Forecasting Engine:\r\n\uf0b7Automated Notification Trigger:$$\nIn BoatMart, Customer Module I am Making following sub modules\r\n1. Conversational AI Commerce Assistant A smart chatbot that helps customers shop like a real person. Users can simply type or speak what they want and the AI understand meaning show the exact matching products.\r\n\uf0b7Natural Language Input Interface: A chat screen where users can type or send voice messages.\r\n\uf0b7Intent Extraction Engine: The AI that picks out important details like color, price, and category from the user's message.\r\n\uf0b7Semantic Search & Matcher: Searches the database using the meaning of the words, not just exact spellings.\r\n2. Automata-Driven Negotiation Engine A smart bargaining bot that lets customers negotiate prices\r\n\uf0b7Negotiation Chat Interface: \r\n\uf0b7State Machine Logic Controller: \r\n\uf0b7Wholesale Margin Calculator: \r\n3. Reinforcement Learning Discovery Pipeline A fun, Tinder-style shopping screen where customers swipe right to like an item or left to skip. \r\n\uf0b7Swiping Card UI: \r\n\uf0b7User Action Logger: \r\n\uf0b7RL Weight Updater:$$\nIn BoatMart, Customer Module I am Making following sub modules\r\n1. Multiplayer Social Commerce Gateway A feature that lets friends shop together at the same time. Users can share a cart link on WhatsApp, add items together, split the bill, and get special bulk discounts from wholesalers for ordering together.\r\n\uf0b7WebSocket Connection Manager: \r\n\uf0b7Order Pooling Engine:.\r\n\uf0b7Wholesale Discount Calculator: Automatically lowers the price when the group orders enough items together.\r\n\uf0b7Bill Split & Checkout Calculator: Divides the total cost so each friend knows exactly what they have to pay.\r\n2. Algorithmic Trust & Sentiment Framework A system designed to build trust by showing honest reviews and reliable sellers. \r\n\uf0b7Sentiment Classification Model:\r\n\uf0b7Seller Metrics Aggregator: \r\n\uf0b7Real-time Score Updater: \r\n3. Predictive Analytics & Forecasting Hub A smart tracking system that guesses when a product's price will go down. \r\n\uf0b7Watchlist Management Interface: \r\n\uf0b7Time-Series Forecasting Engine: \r\n\uf0b7Automated Notification Trigger:$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-18 22:58:20","updated_at":"2026-03-18 22:58:20","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1093,"project_id":1473,"title":"BOATMART - Module C \r\nThe All-in-One Solution for E-Commerce","prob":"The e-commerce industry is highly fragmented: wholesalers, retailers, and dropshippers operate on separate platforms and rely on multiple disconnected tools for inventory, orders, customer management, payments, analytics, CRM, and ERP. This lack of integration increases operational costs, creates data inconsistency, slows workflows, and adds technical barriers\u2014especially for small and medium businesses\u2014making it harder to scale.\r\nThis FYP proposes a lightweight all-in-one e-commerce prototype that unifies key operations in a single system by connecting wholesalers, retailers, and dropshippers. The platform will centralize a shared product catalog, basic inventory visibility, order processing and routing, and simple status updates, along with supportive built-in tools such as basic CRM, sales analytics dashboards, reporting, and automation features. By reducing dependence on multiple tools, the solution aims to simplify workflows, improve efficiency, provide clearer data visibility, and lower operational complexity for growing businesses","description":"The proposed system is a unified e-commerce ecosystem designed to connect wholesalers, dropshippers\/sellers, and retail customers within a single integrated platform, eliminating the need for multiple tools and platforms.\r\nWholesalers register on the platform and access a dedicated supplier dashboard where they upload and manage product information, including pricing, variants, available stock, and estimated fulfillment times. Wholesalers do not sell directly to end customers; instead, they act as product suppliers for sellers on the platform.\r\nDropshippers or sellers create their own branded storefronts within the system. They can browse the unified product catalog, select products from registered wholesalers, and list them in their stores with customized pricing, branding, and product descriptions.\r\nRetail customers browse seller storefronts or a centralized marketplace, place orders, and track deliveries. Once an order is placed, the platform automatically routes it to the relevant wholesaler for fulfillment, manages order tracking, and handles secure payment distribution among the wholesaler, seller, and platform.\r\nThe system implements role-based authentication and authorization (customer, seller, wholesaler, administrator), ensuring secure access and controlled functionality for each user type.$$\nC1. Automated Supplier Verification & Onboarding Engine\r\n\r\nManages secure onboarding of wholesalers through automated validation.\r\nSubmodules\r\n\u2022 Document Submission Interface \u2013 Upload CNIC, NTN, trade license.\r\n\u2022 OCR & Data Extraction \u2013 Extracts data from documents.\r\n\u2022 Document Validation Engine \u2013 Ensures format and data consistency.\r\n\u2022 Fraud Detection Check \u2013 Detects suspicious or altered documents.\r\n\u2022 Blacklist Screening \u2013 Matches suppliers with restricted records.\r\n\u2022 Auto Approval System \u2013 Generates score and approves\/rejects.\r\n\r\nC2. AI-Assisted Product Catalog & Data Ingestion System\r\n\r\nEnables efficient product listing using AI automation.\r\nSubmodules\r\n\u2022 Manual Product Upload Interface \u2013 Add products with images.\r\n\u2022 AI Image Refinement Module \u2013 Enhances product visuals.\r\n\u2022 AI Product Description Generator \u2013 Creates titles and descriptions.\r\n\u2022 Bulk Catalog Import Engine \u2013 Upload via Excel\/CSV.\r\n\u2022 AI Field Mapping System \u2013 Detects column mapping.\r\n\u2022 Data Validation & Error Reporting \u2013 Flags missing\/duplicate data.\r\n\u2022 Batch Image Processing Module \u2013 Processes bulk images.\r\n\r\nC3. Smart Inventory & Multi-Warehouse Management System\r\n\r\nMaintains real-time inventory across warehouses.\r\nSubmodules\r\n\u2022 Inventory Ledger System \u2013 Tracks stock movement.\r\n\u2022 Multi-Warehouse Management \u2013 Monitors multiple locations.\r\n\u2022 Low Stock Alert System \u2013 Sends stock alerts.\r\n\u2022 Stock Reconciliation Tool \u2013 Matches physical vs system stock.\r\n\u2022 Sell-Rate Monitoring Engine \u2013 Estimates product demand rate.\r\n\r\nC4. Intelligent Order Processing & Fulfillment Engine\r\n\r\nHandles complete order lifecycle management.\r\nSubmodules\r\n\u2022 Order Reception Dashboard \u2013 Displays incoming orders.\r\n\u2022 Order Status Workflow Manager \u2013 Tracks order stages.\r\n\u2022 Inventory Reservation System \u2013 Reserves stock on confirmation.\r\n\u2022 Dispute Handling Module \u2013 Manages complaints\/issues.\r\n\u2022 Fulfillment Monitoring System \u2013 Tracks progress and delays.\r\n\r\nC5. Integrated Logistics & Delivery Orchestration System\r\n\r\nAutomates shipping and delivery operations.\r\nSubmodules\r\n\u2022 Courier Integration Setup \u2013 Connect courier services.\r\n\u2022 Automatic Shipment Booking \u2013 Schedules pickups.\r\n\u2022 Smart Courier Selection Engine \u2013 Chooses optimal courier.\r\n\u2022 Shipping Label Generator \u2013 Creates labels.\r\n\u2022 Real-Time Shipment Tracking \u2013 Updates delivery status.\r\n\u2022 Delivery Confirmation Handler \u2013 Marks completed orders.\r\n\r\nC6. Business Intelligence & Operational Analytics Engine\r\n\r\nProvides insights through reports and dashboards.\r\nSubmodules\r\n\u2022 Sales Performance Dashboard \u2013 Tracks revenue trends.\r\n\u2022 Product Performance Analyzer \u2013 Identifies top\/slow products.\r\n\u2022 Inventory Health Report \u2013 Shows stock risks.\r\n\u2022 Fulfillment Efficiency Report \u2013 Measures delivery performance.\r\n\u2022 Report Export Module \u2013 Generates downloadable reports.\r\n\r\nC7. Workflow Automation & Smart Notification System\r\n\r\nAutomates tasks and enables real-time communication.\r\nSubmodules\r\n\u2022 Automation Rule Engine \u2013 Defines if-then rules.\r\n\u2022 Auto Order Confirmation \u2013 Confirms trusted orders.\r\n\u2022 Task Assignment System \u2013 Assigns warehouse tasks.\r\n\u2022 Real-Time Notification Service \u2013 Sends alerts\/updates.\r\n\u2022 Escalation Alert Module \u2013 Flags delays\/issues.\r\nC8. Autonomous Supply Chain Intelligence Engine\r\nProvides predictive insights for supply chain optimization.\r\nSubmodules\r\n\u2022 Demand Forecasting Engine \u2013 Predicts future demand.\r\n\u2022 Inventory Risk Analyzer \u2013 Detects stock shortages.\r\n\u2022 Demand Spike Detection System \u2013 Identifies demand surges.\r\n\u2022 Fulfillment Recommendation Engine \u2013 Suggests optimal logistics.\r\n\u2022 Slow-Moving Product Detector \u2013 Identifies low-performing items.\r\n\u2022 Supply Chain Alert System \u2013 Generates proactive alerts .$$\nMember 1 \u2013 Pre-Order Supply Management\r\nResponsible for functionalities that operate before an order is placed and for maintaining the supplier\u2019s operational readiness, product catalog integrity, inventory visibility, and analytical insights for decision-making.\r\nModules\r\nC1 \u2014 Automated Supplier Verification & Onboarding Engine\r\nC2 \u2014 AI-Assisted Product Catalog & Data Ingestion System\r\nC3 \u2014 Smart Inventory & Multi-Warehouse Management System\r\nC6 \u2014 Business Intelligence & Operational Analytics Engine\r\nFocus Area\r\nSupplier onboarding, AI-assisted product catalog management, warehouse inventory tracking, stock health monitoring, and analytical reporting. This member ensures the system has verified suppliers, structured product data, accurate inventory levels, and actionable business insights before orders are placed.$$\nMember 2 \u2013 Post-Order Fulfillment & Supply Intelligence\r\nResponsible for functionalities that operate after seller orders are received, focusing on order lifecycle management, logistics coordination, operational automation, and supply chain intelligence.\r\nModules\r\nC4 \u2014Intelligent Order Processing & Fulfillment Engine\r\nC5 \u2014 Integrated Logistics & Delivery Orchestration System\r\nC7 \u2014Workflow Automation & Smart Notification System\r\nC8 \u2014Autonomous Supply Chain Intelligence Engine\r\nFocus Area\r\nOrder processing, shipment coordination, delivery tracking, operational automation, and predictive supply chain intelligence to ensure efficient order fulfillment and proactive supply management.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-18 22:59:03","updated_at":"2026-03-18 22:59:03","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1090,"project_id":1455,"title":"AI-Enhanced Blockchain-Based Smart Rental System for Cameras & Photography Equipment","prob":"Professional photography equipment such as cameras, lenses, drones, and gimbals is expensive and not easily affordable for students and freelancers. Existing rental systems rely on manual agreements and lack transparency, which leads to disputes regarding payments, late returns, and equipment damage.\r\nMoreover, current systems do not support intelligent pricing, automated damage detection, or real-time availability tracking. Blockchain solutions used in similar systems are often limited to wallet-based transactions (e.g., MetaMask) and lack scalability and advanced automation.\r\nThis project aims to develop an AI-powered blockchain-based rental platform that ensures transparency, automation, smart pricing, damage detection, and efficient real-time rental management.","description":"The system is a web-based + AI-integrated rental platform where users can rent cameras and photography equipment securely.\r\nEquipment owners can list their items, while renters can search, book, and rent items based on availability. A smart contract is automatically generated for each rental, defining payment, deposit, duration, and penalties.\r\nThe system is enhanced with:\r\nAI-based dynamic pricing\r\nImage-based damage detection\r\nAdvanced smart contracts with conditional execution\r\nBlockchain ensures transparency, while AI improves automation and decision-making. The system also handles real-time availability updates, late returns, and misuse scenarios efficiently, making it more advanced than a typical e-commerce platform.$$\n1. User Management Module \r\n2. Equipment Management Module \r\n3. Rental & Booking Module \r\n4. Blockchain & Smart Contract Module \r\n5. Payment Module\r\n6. AI-Based Dynamic Pricing Module\r\n7. Image-Based Damage Detection Module\r\n8. Real-Time Monitoring & Availability Module$$\nI will develop the Equipment Management Module, Payment Module, Image-Based Damage Detection Module, and Real-Time Monitoring & Availability Module. This includes managing equipment listings, handling secure payments, detecting damages using image comparison techniques, and ensuring real-time tracking of equipment availability to prevent booking conflicts.$$\nI will develop the User Management Module, Rental & Booking Module, Blockchain & Smart Contract Module, and AI-Based Dynamic Pricing Module. This includes handling user authentication, managing bookings, implementing smart contracts for secure transactions, and developing an AI model to dynamically adjust rental prices based on demand and usage trends.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-17 20:00:51","updated_at":"2026-03-17 20:00:51","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1078,"project_id":1490,"title":"Cloud Native ERP & HRMS: VAT Compliant Invoicing, Inventory, and Payroll Automation for Businesses.","prob":"Small and Medium Enterprises in the UAE currently struggle with fragmented operations using separate unconnected software for accounting, inventory, and HR. This leads to data redundancy and errors. Specifically, manual calculation of UAEs\u2019 5% VAT and the new Corporate Tax is prone to mistakes leading to audit risks. Furthermore, manual tracking of employee documents like Emirates ID, Labour Cards, Visa\u2019s, results in missed renewal deadlines and significant government fines. Existing ERPs are either too expensive or lack specific local compliance features like Payroll and tax laws, forcing businesses to rely on inefficient spreadsheets.","description":"This project is a unified Enterprise Resource Planning (ERP) and Human Resource Management system designed specifically for the UAE market. It streamlines business operations by integrating three critical domains: Finance, Inventory, and HR.\r\n1. Financial & Tax Management: The system allows users to create tax compliant invoices and quotations. It automatically calculates VAT (5%) on sales and purchases. The core feature is the Tax Engine which generates reports ready for UAE Federal Tax Authority (FTA) filing, including Corporate Tax provisions.\r\n2. Inventory & Operations: It provides real time tracking of stock levels across multiple warehouse locations. The system supports Low Stock Alerts and automates the creation of Purchase Orders (POs) to suppliers when inventory hits a threshold. It also tracks project specific expenses to calculate accurate profitability.\r\n3. HR & Payroll Automation: The system digitizes employee records and includes a Document Watchdog that tracks expiry dates for Passports, Visas, and Emirates IDs, sending alerts to HR before expiry to prevent fines. The payroll module processes salaries.\r\n4. Mobile Accessibility: An Employee Self-Service mobile application allows employees to view pay slips, work hours and monthly attendance.$$\nIn the Cloud Native ERP & HRMS System, the modules are: (i) Financial Accounting & Tax (Invoicing, VAT, Expenses), (ii) Inventory Management (Stock tracking, Purchase Orders), (iii) HR Management (Employee records, Visa\/ID tracking), (iv) Payroll Automation (Salary calculation, WPS compliance), and (v) Employee Mobile App (Self-service portal for attendance and leave)$$\nI will develop Business Operations Module (Finance & Inventory). My tasks include: (i) Creating the Invoicing system that automatically calculates 5% VAT and Corporate Tax, (ii) Building the Tax Engine to generate reports for the government, (iii) Developing the Inventory system to track stock levels and alert on low stock, and (iv) Managing Purchase Orders and suppliers.$$\nI will develop the HR Management & Mobile App Module. My tasks include: (i) Building the Employee Database with the \"Document Watchdog\" to track Visa\/ID expiry dates, (ii) Creating the Payroll system to calculate salaries and overtime according to UAE laws, (iii) Developing the Attendance logic, and (iv) Building the Mobile App (React Native\/Flutter) for employees to view payslips and mark attendance.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-14 23:49:01","updated_at":"2026-03-14 23:49:01","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1075,"project_id":1498,"title":"Story Nest: an Agentic AI-Based Interactive Storytelling Application for Children","prob":"Many existing digital storytelling applications provide pre-defined or repetitive story content with limited interaction. Such applications do not adapt stories according to a child\u2019s age, interests, or engagement level, which can reduce attention and interest over time. Parents also face difficulty finding storytelling platforms that are both engaging and suitable for children. As a result, children may lose interest in storytelling activities, limiting the educational and creative benefits that storytelling can offer. There is a need for a more interactive and adaptive storytelling solution that can provide engaging and appropriate story experiences for children.","description":"This project focuses on the development of an Agentic AI-based interactive storytelling mobile application for children. The application aims to provide an engaging storytelling experience by generating stories dynamically based on user input. An agentic AI-based backend is used to manage the storytelling process, where different components of the system work together to create structured and suitable story content. The application combines story text with visual and audio elements to enhance engagement and maintain children\u2019s interest. The system is designed to be child-friendly, interactive, and adaptable, providing an enjoyable digital storytelling environment. In this project, students will implement the agentic storytelling workflow, structured prompt engineering for story generation, and branching logic for interactive decision-making. Students will also implement context management to maintain story continuity, character voice narration using neural Text-to-Speech integration, and content safety mechanisms to ensure age-appropriate storytelling. Additionally, the backend will coordinate multiple AI services to ensure controlled and consistent story progression.$$\n1) User Profile & Preferences Module\r\nManages basic user information and preferences required for personalized storytelling.\r\n2) AI-Based Story Generation Module\r\nStudents will implement generative AI integration to create age-appropriate and engaging story segments based on user inputs such as theme, preferences, and choices. This module will use structured prompts to guide the AI in generating consistent and dynamic story content as the story progresses. Since story generation relies on external APIs, delays may occur because the requests are processed on remote cloud servers and depend on factors such as network latency, server load, and the complexity of generated content. To reduce this issue, the system will use asynchronous API requests, loading indicators, caching of previously generated content, and optimized prompts to improve response time and maintain a smooth user experience.\r\n3) Agentic Story Management Module\r\nCoordinates different AI components to manage the storytelling workflow in a structured manner.\r\n4) Character Voice Handling Module\r\nManages narration and character-specific voices during story playback. This module integrates Microsoft Azure Speech Text-to-Speech to convert generated story text into audio. Azure\u2019s neural voices will be evaluated for naturalness and speech quality. Emotional expression in narration will be enhanced either through SSML controls (such as pitch, rate, and speaking style) to adjust intonation and connotations, or by using pre-recorded sound effects that match the scene context. Selected voices will then be mapped to roles such as narrator and characters to improve the overall storytelling experience.\r\n5) Story Presentation & Interaction Module\r\nDisplays the generated story in the mobile application along with audio narration and visual elements to create an engaging storytelling experience. The module also supports interactive storytelling by presenting the user with decision options at specific points in the story. When a user selects an option, the choice is sent to the backend, which uses the previous story context to generate the next segment of the story. This allows the narrative to branch and evolve dynamically based on user decisions, making each storytelling session unique and interactive.\r\n6) Backend & Data Management Module\r\nManages data storage, retrieval, and communication between the application and backend services.$$\n1) AI-Based Story Generation Module:\r\nDesigns and implements intelligent story creation using generative AI to produce age-appropriate, engaging, and dynamic stories based on user inputs and preferences.\r\n2) Agentic Story Management Module:\r\n Handles autonomous decision-making, story flow control, and character actions to ensure logical progression and adaptive storytelling.\r\n3) Backend & Data Management Module: \r\nManages databases, APIs, authentication, and system integration to ensure secure storage, efficient data retrieval, and smooth communication between application components.$$\n1) User Profile & Preferences Module:\r\n Develops user accounts, manages child profiles, and stores preferences such as age, interests, and language for personalized story experiences.\r\n2) Story Presentation & Interaction Module: \r\nDesigns interactive user interfaces, story visualization, and navigation to enhance engagement and usability for children.\r\n3) Character Voice Handling Module: \r\nImplements voice selection, audio synchronization, and text-to-speech integration to deliver expressive and immersive character narration.$$\n$$\nQissa Sunao: An Urdu Storytelling Platform Powered by AI$$\n1. Preference-Based Story Creation$$\n2. Agentic AI-Based Story Management$$\n3. Multi-Voice Storytelling Experience","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-14 14:26:54","updated_at":"2026-03-14 14:26:54","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1082,"project_id":1451,"title":"An Integrated Multimodal System for Detecting AI-Generated and Manipulated Media","prob":"The rapid advancement of generative artificial intelligence has made it increasingly difficult to distinguish between real and artificially generated or manipulated digital content. Deepfake videos, AI-generated images, synthetic audio, and machine-generated text are being misused for misinformation, fraud, identity impersonation, fake news, and social manipulation. Currently, detection tools are fragmented, media-specific, and require technical expertise, making them impractical for general users and organizations. There is no single, unified platform that can analyze multiple content types\u2014image, video, audio, and text\u2014within one application. This creates a critical gap where individuals, journalists, educators, and institutions lack accessible and reliable means to verify content authenticity. The proposed FYP software addresses this problem by providing an integrated system that automatically detects whether digital content is AI-generated, manipulated, or authentic, enabling timely verification and helping reduce the spread of deceptive and harmful digital media.","description":"The proposed FYP is a multimodal content verification application designed to detect AI-generated, manipulated, and synthetic media across multiple formats, including images, videos, audio, and text. The application provides a unified platform where users can upload digital content and receive an automated authenticity analysis. After user authentication, the system allows registered users to submit media files or text through a simple interface.\r\n\r\nFor image and video inputs, the system analyzes visual artifacts, inconsistencies, and deepfake indicators using machine learning\u2013based detection models. Audio inputs are examined for synthetic speech patterns, unnatural frequency variations, and voice generation artifacts to determine whether the audio is real or AI-generated. Text inputs are analyzed using linguistic and statistical features to identify AI-generated writing patterns. Each detection module operates independently but is integrated into a single backend system, enabling multimodal analysis within one application.\r\n\r\nOnce processing is complete, the application generates a detailed result indicating whether the content is authentic, AI-generated, or manipulated, along with a confidence score. The system also stores analysis history, allowing users to review previously checked content. An administrative panel is provided to manage users, monitor system usage, and update detection models. The application is designed to be scalable and user-friendly, making advanced content verification accessible to non-technical users.\r\n\r\nBy integrating multiple detection capabilities into one platform, the proposed system eliminates the need for separate tools for different media types and provides a practical solution for combating misinformation, fraud, and digital impersonation in real-world scenarios.$$\nThe proposed system is divided into the following major modules to ensure modularity, scalability, and efficient content analysis:\r\n\r\n1. User Authentication and Authorization Module\r\n\r\nThis module handles user registration, login, and logout functionalities. It ensures secure access to the system using role-based authorization. Registered users can upload and analyze content, while administrators can manage users and system configurations. This module also maintains user sessions and protects the application from unauthorized access.\r\n\r\n2. Image Manipulation Detection Module\r\n\r\nThis module is responsible for analyzing uploaded images to determine whether they are AI-generated or manipulated. It examines visual inconsistencies, artifacts, and generative patterns using trained machine learning models. The module processes images in supported formats and returns an authenticity classification along with a confidence score.\r\n\r\n3. Video Deepfake Detection Module\r\n\r\nThis module analyzes video content to detect deepfakes and other forms of video manipulation. It extracts key frames and temporal features to identify facial distortions, unnatural motion patterns, and frame-level inconsistencies. The system outputs whether the video is authentic or manipulated.\r\n\r\n4. Audio Authenticity Detection Module\r\n\r\nThe audio module detects whether uploaded audio files are real or synthetically generated. It analyzes frequency patterns, pitch variations, and spectral features commonly found in AI-generated speech. The module supports common audio formats and provides a clear authenticity result with confidence metrics.\r\n\r\n5. Text AI-Generation Detection Module\r\n\r\nThis module evaluates textual input to determine whether it is human-written or AI-generated. It uses linguistic, statistical, and stylistic features such as sentence structure, repetition, and coherence patterns. The module is suitable for analyzing articles, posts, or short textual content.\r\n\r\n6. Result Generation and Reporting Module\r\n\r\nThis module aggregates outputs from all detection modules and presents results in a user-friendly format. It displays the classification result, confidence score, and media type. Users can download or view reports of their analysis.\r\n\r\n7. History and Data Management Module\r\n\r\nThis module stores analysis records, including uploaded content metadata, results, and timestamps. Users can view their previous analyses, while administrators can monitor system usage for performance and auditing purposes.\r\n\r\n8. Admin Management Module\r\n\r\nThe admin module allows system administrators to manage users, view logs, update detection models, and monitor overall system health. It ensures smooth operation and future scalability of the application.$$\nI shall develop the core administrative and text-analysis modules, which include:\r\n\r\n1.User Authentication and Authorization Module: To ensure secure access and data privacy for all users.\r\n\r\n2.Result Generation and Reporting Module: To compile detection data into comprehensive and exportable reports.\r\n\r\n3.History and Data Management Module: For storing and retrieving past analysis results efficiently.\r\n\r\n4.Text AI-Generation Detection Module: To identify synthetically generated text using natural language processing.\r\n\r\n5.Admin Management Module: For overseeing platform operations and user roles$$\nI shall develop the specialized multimedia forensic modules, which include:\r\n\r\n1.Image Manipulation Detection Module: To identify edited or doctored images by analyzing pixel inconsistencies and metadata.\r\n\r\n2.Video Deepfake Detection Module: To detect synthetic facial swaps and temporal artifacts in video content using deep learning models.\r\n\r\n3.Audio Authenticity Detection Module: To differentiate between genuine human speech and AI-cloned or synthesized audio tracks to ensure overall media integrity.$$\n$$\nDeepShield: Protecting Against Deepfake Manipulations$$\n1. Audio Authenticity Detection Module$$\n2. Text AI-Generation Detection Module$$\n3. Multimodel support for detecting different types of image\/video content","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-15 20:27:18","updated_at":"2026-03-24 12:53:38","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1080,"project_id":1465,"title":"EntangledEcho: Quantum-Inspired AI Social Media Persona Optimizer for Creators","prob":"Small creators and businesses struggle to grow on social media due to inefficient content\r\nplanning, random posting schedules, and low engagement. They often spend hours guessing\r\nwhat to post, when to post, and which hashtags or formats will work, resulting in wasted\r\neffort, slow audience growth, and missed monetization opportunities. EntangledEcho\r\nprovides AI-driven, data-backed recommendations to optimize content strategy, reduce\r\nmanual planning, and increase online engagement and reach.","description":"EntangledEcho is a web-based platform that combines AI and quantum-inspired\r\noptimization to help creators grow efficiently on social media. The system analyzes\r\nhistorical engagement data, audience behavior, and content types, then predicts the optimal\r\nposts, timing, and hashtags for maximum reach. Users receive actionable recommendations,\r\nvisual analytics, and automated notifications. The platform reduces manual planning by 60\u2013\r\n70% while enabling smarter, personalized posting strategies. Quantum-inspired\r\noptimization ensures the most effective content combinations are suggested without\r\nrequiring actual quantum hardware.$$\nIn EntangledEcho, the modules are (i)User Management & Authentication (ii)Data Collection Module (iii)Content Analysis Module (iv)Quantum-Inspired Optimization Module (v)Recommendation & Analytics Module (vi)AI Task Decomposition (vii)Security & Privacy Module (viii)Content Scheduling & Automation Module (ix)Social Media Insights Module (x)Notification Module$$\nIn EntangledEcho, i shall develop (ii)Data Collection Module Responsible for gathering engagement metrics from connected social media accounts, such\r\nas likes, comments, shares, and views. It normalizes the collected data and stores it\r\nefficiently for analysis.(iii)Content Analysis Module that Analyzes posts, captions, hashtags, media types, and posting times to generate engagement\r\nscores and identify content performance patterns. (iv)Quantum-Inspired Optimization Module that Uses quantum-inspired algorithms to suggest optimal content schedules and combinations.\r\nIt simulates multiple posting strategies to predict the highest engagement potential. (viii)Content Scheduling & Automation Module Enables automatic scheduling of posts based on AI recommendations and integrates with\r\nsocial media APIs for seamless posting. (ix)Social Media Insights Module that Displays audience demographics, engagement trends, and competitor analysis, helping\r\nusers make informed decisions.$$\nIn EntangledEcho, i shall develop (i)User Management & Authentication his module manages the entire user lifecycle, including registration, login, and profilemanagement. It supports OAuth integration for linking social media accounts and implements\r\nrole-based access control to differentiate permissions for Creators, Analysts, and Admins. (vi)AI Task Decomposition Breaks down content strategies into actionable tasks such as micro-posts, story ideas, or\r\ncampaigns. Users can edit AI-generated suggestions before posting. (v)Recommendation & Analytics Module Breaks down content strategies into actionable tasks such as micro-posts, story ideas, or\r\ncampaigns. Users can edit AI-generated suggestions before posting. (vii)Security & Privacy Module Ensures secure storage of user data and OAuth tokens through encryption, access control,\r\nand secure API communication. (x)Notification Module Sends alerts for optimal posting times, content reminders, and engagement updates to\r\nensure timely actions.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-15 12:18:36","updated_at":"2026-03-15 12:18:36","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1089,"project_id":1456,"title":"CloudSight: Intelligent Cloud Data Protection & Governance","prob":"Organizations today suffer from \"Data Blindness,\" accumulating vast amounts\r\nof unstructured documents (contracts, reports) on cloud storage that remain\r\nunsearchable beyond basic filenames. While AI offers the potential to unlock\r\nthis data, it introduces a critical Privacy Paradox: businesses cannot safely\r\nfeed sensitive internal files into public Large Language Models (LLMs) without\r\nrisking severe data leaks and regulatory non-compliance (e.g., exposing CNICs\r\nor financial data). Furthermore, the operational cost and latency of processing\r\nentire cloud drives via commercial APIs are impractical. Currently, there is no\r\nunified solution that allows organizations to intelligently query their private\r\ndata while enforcing strict, local-first privacy governance to prevent sensitive\r\ninformation from ever leaving their control.","description":"CloudSight transforms standard cloud storage into a secure, intelligent\r\nknowledge base. The system connects to Google Drive via OAuth 2.0, utilizing\r\nan efficient incremental sync engine to fetch and monitor files. To address\r\ncritical privacy risks, it employs a proprietary \"Hybrid Privacy Scanner.\"\r\nIncoming documents are first processed locally using Regex and lightweight\r\nNLP models to detect and redact sensitive PII (such as CNICs or credit card\r\nnumbers) before any data is sent to external AI providers. This ensures strict data\r\nsovereignty and significantly reduces API costs.\r\nOnce secured, the data is indexed using a \"Hybrid Search\" architecture\r\n: Elasticsearch handles metadata and exact keyword matching, while a Vector\r\nDatabase captures semantic meaning. This infrastructure powers evidence\r\n-based chatbot (RAG) that allows users to query their private documents in\r\nnatural language. Unlike standard AI, the system provides \"trustworthy\"\r\nanswers by including clickable citations that link directly to the specific source\r\nparagraph in the original file.$$\nIn CloudSight, the modules are (i) Secure Cloud Connector & Sync Engine, (ii) Hybrid\r\nPrivacy & Governance Scanner, (iii) Asynchronous Data Processing Pipeline, (iv)\r\nAdministrative Command Center & Metadata Store, (v) Intelligent Document Processor\r\n& Vectorization, (vi) Hybrid Search & Retrieval Core, (vii) Context-Aware Conversational\r\nInterface, and (viii) Citation & Source Verification Layer.$$\nIn CloudSight, I shall develop the Governance & Ingestion modules which\r\ninclude: (i) Secure Cloud Connector (OAuth 2.0 authentication & Incremental\r\nSyncing), (ii) Hybrid Privacy Scanner (Local Tier-1 PII detection & automated\r\nredaction), (iii) Asynchronous Data Pipeline (Redis-based queue for non\r\n-blocking ingestion), and (iv) Administrative Command Center (Real-time\r\nactivity monitoring & Elasticsearch metadata indexing).$$\nIn CloudSight, I shall develop the Intelligence & Retrieval modules which include: (i)\r\nIntelligent Document Processor (Semantic text chunking & Vector embedding\r\ngeneration), (ii) Hybrid Search Core (Combining keyword & vector search algorithms),\r\n(iii) Context-Aware Conversational Interface (LLM-powered chatbot with query\r\nmemory), and (iv) Citation Verification Layer (Logic to map AI responses back to\r\noriginal source documents).$$\n$$\n$$\n1. OCR Vision Engine (Image-Based PII Redaction)\r\nIntegrates Tesseract OCR to extract and redact PII from physical or scanned documents (JPG, PNG) in Google Drive before AI vectorization.$$\n2. Dynamic Forensic Watermarking (Insider Threat Mitigation)\r\nOverlays a dynamic watermark (email, IP, time) on viewed files to stop insider data exfiltration via screen captures.$$\n3. Cross-Lingual Semantic Retrieval (Bilingual Support)\r\nEnables Roman\/Native Urdu queries. The AI translates inputs, searches the English database, and replies in Urdu.","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-17 02:45:10","updated_at":"2026-03-17 03:54:51","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}
No. Project Title Reg. No. Name Date & Time Evaluation Status
1. Playdex CIIT/SP23-BCS-020/WAH
CIIT/SP23-BCS-040/WAH
 NAVAIRA
HAMZA ALI
 Done Accepted  
2. Intelligent RFP & Compliance Agent: AI Powered Questionnaire Automation with Format Preservation CIIT/SP23-BSE-006/WAH
CIIT/SP23-BSE-038/WAH
 MUHAMMAD ZUBAIR
FARHAD ALI TAIMOOR
 Done Accepted  
3. Cooperative Multiplayer Escape Game CIIT/SP23-BSE-035/WAH
CIIT/SP23-BSE-060/WAH
 MUSTAFA RIAZ
MOHEES AHMED
 Done Accepted  
4. AR-Space Planner: A Real-Time Augmented Reality System for Interior Design and Furniture Visualization CIIT/SP23-BSE-024/WAH
CIIT/SP23-BSE-044/WAH
 RIZWAN HABIB
BUSHRA TANVEER
 Feb 13, 2026
02:00 PM 		Pending  
5. HEMS – Hybrid Emergency Messaging System with Resilient Mesh Networking and Adaptive Relay Management CIIT/SP23-BCS-058/WAH
CIIT/SP23-BCS-100/WAH
 UM-E-FARWA RIZVI
HAMZA IFTIKHAR
 Feb 13, 2026
02:20 PM 		Pending  
6. SafeToon: AI-Powered Parental Control and Smart Cartoon Recommendation System for Children on YouTube. CIIT/SP23-BSE-011/WAH
CIIT/SP23-BSE-015/WAH
 TEHREEM BATOOL
HASSAAN ABDULLAH
 Feb 13, 2026
02:40 PM 		Revision Pending 
7. Quit-Companion: A mobile app supporting nicotine recovery through habit tracking, personalized guidance, improved overall health. CIIT/SP23-BCS-071/WAH
CIIT/SP23-BCS-084/WAH
 SUKAINA AYESHA
MEHREEN ALI
 Feb 13, 2026
03:40 PM 		Pending  
8. AI-Assisted Augmented Reality System for Product Size and Fit Visualization in Online Shopping CIIT/SP23-BCS-064/WAH
CIIT/SP22-BCS-050/WAH
 AYESHA SHAKEEL
Sehar Siddique
 Feb 23, 2026
08:05 AM 		Pending  
9. MindGuard: AI-Powered Mental Well-being Monitoring and Insight System CIIT/SP23-BCS-062/WAH
CIIT/SP23-BCS-089/WAH
 ARMNAN ALI
ALI SHAZIL
 Feb 23, 2026
08:45 AM 		Revision Pending 
10. Skill-Nest – An AI-powered platform that grows your skills and guides to perfect job CIIT/SP23-BSE-012/WAH
CIIT/SP23-BSE-023/WAH
 MUHAMMAD AFZAAL HASSAN
AHMER ABBAS
 Feb 23, 2026
09:10 AM 		Revision Pending 
11. Web-Based Enterprise Resource Planning (ERP) System for Young’s Food Pvt Ltd CIIT/SP23-BSE-008/WAH
CIIT/SP23-BSE-054/WAH
 ADNAN HAIDER
AHMAR ISHAQ
 Feb 23, 2026
09:25 AM 		Revision Pending 
12. NurtureAI: AI-Powered Maternal Mental Health Support Chatbot for for Pre- and Postnatal Mothers CIIT/SP23-BCS-013/WAH
CIIT/SP23-BCS-102/WAH
 OMAIMA KHALID
ROOMAH FATIMA BUKHARI
 Feb 25, 2026
09:00 AM 		Revision Pending 
13. Vendra: A Multi-Vendor Digital Marketplace for Global Commerce CIIT/SP23-BCS-045/WAH
CIIT/SP23-BCS-088/WAH
 MUHAMMAD BILAL
ZOHAIB KHAN
 Feb 25, 2026
09:25 AM 		Revision Pending 
14. CampusShed AI – Next‑Gen Scheduling (Industrial Project) CIIT/SP23-BSE-021/WAH
CIIT/SP23-BSE-032/WAH
 MUHAMMAD HAMZA
MUNTAHA WANI
 Feb 25, 2026
09:45 AM 		Revision Pending 
15. BOATMART – The All-in-One Solution for E-Commerce Module A CIIT/SP23-BCS-055/WAH
CIIT/SP23-BCS-073/WAH
 SADAF RAMZAN
AHMED ABDULLAH
 Feb 25, 2026
11:05 AM 		Revision Pending 
16. BOATMART - Module C The All-in-One Solution for E-Commerce CIIT/SP23-BCS-053/WAH
CIIT/SP23-BCS-056/WAH
 HUNAIN AHMED
SHUMAISA ASIF
 Feb 25, 2026
11:25 AM 		Revision Pending 
17. AI-Enhanced Blockchain-Based Smart Rental System for Cameras & Photography Equipment CIIT/SP23-BCS-003/WAH
CIIT/SP23-BCS-083/WAH
 SAFINA
ALISHBA MALIK
 Feb 25, 2026
12:00 PM 		Revision Pending 
18. Cloud Native ERP & HRMS: VAT Compliant Invoicing, Inventory, and Payroll Automation for Businesses. CIIT/SP22-BSE-016/WAH
CIIT/SP22-BSE-062/WAH
 Hassan Ul Banna
Muhammad Rumman Khan
 Mar 4, 2026
09:00 AM 		Revision Pending 
19. TubeTrendAI: An AI-Powered YouTube SEO, Trending Analysis and Content Automation Platform CIIT/SP23-BCS-024/WAH
CIIT/SP23-BCS-087/WAH
 HURRARAH QASWER
MOIZ AHMED
 Mar 4, 2026
09:25 AM 		Revision Pending 
20. TrustGate: An Intelligent Access and Governance Platform for Smart Communities CIIT/SP23-BCS-010/WAH
CIIT/SP23-BCS-097/WAH
 ZULAIKHA BATOOL
AMMARA GUL
 Mar 4, 2026
09:45 AM 		Revision Pending 
21. FIT STATE HUB CIIT/SP23-BCS-079/WAH
CIIT/FA22-BSE-006/WAH
 HAFIZA HIFZA MAQBOOL
Muhammad Momin
 Mar 4, 2026
11:05 AM 		Revision Pending 
22. ZaraiLink: A Multi-Vendor Agribusiness Platform with Inventory Management and AI Crop Disease Advisory CIIT/SP23-BCS-046/WAH
CIIT/SP23-BCS-106/WAH
 TARIQ HUSSAIN
HASNAT AFZAL
 Mar 4, 2026
11:25 AM 		Revision Pending 
23. Story Nest: an Agentic AI-Based Interactive Storytelling Application for Children CIIT/SP23-BSE-001/WAH
CIIT/SP23-BSE-030/WAH
 UMER JEHANZEB KHAN
SARIM IQBAL
 Mar 4, 2026
11:50 AM 		Revision Pending 
24. An Integrated Multimodal System for Detecting AI-Generated and Manipulated Media CIIT/SP23-BSE-009/WAH
CIIT/SP23-BSE-016/WAH
 ADEEL SHAREEF
SYED MUHAMMAD HAMZA HASSAN
 Mar 4, 2026
12:05 PM 		Pending  
25. EntangledEcho: Quantum-Inspired AI Social Media Persona Optimizer for Creators CIIT/SP23-BSE-047/WAH
CIIT/SP23-BSE-055/WAH
 ABDUL HAFEEZ
UMER JALAL CHAUDHRY
 Mar 4, 2026
12:25 PM 		Revision Pending 
26. CloudSight: Intelligent Cloud Data Protection & Governance CIIT/SP23-BCS-044/WAH
CIIT/SP23-BCS-107/WAH
 MUHAMMAD ALI SIDDIQUI
AZAN JAHAN
 Mar 4, 2026
12:35 PM 		Pending  
Committee: 2
Team Members:
Dr. Tassawar Iqbal (HEAD)
Dr. Samia Riaz
Mr. Muhammad Nadeem
Mr. Hassan Sardar
Venue: New CS Conference Room
Remarks: All students are advised to reach venue on time for presentation. In case of any issue please contact 0312-0877225

1{"id":1054,"project_id":1476,"title":"SmartStudyInstructor: AI-Powered Hybrid Courses Platform with Adaptive, Personalized, Interactive Lecture Videos, Attendance, Engagement, and Assessments","prob":"Universities increasingly offer hybrid learning, yet existing platforms lack an integrated mechanism to convert teacher materials into structured lecture delivery with measurable participation and learning analytics. Teachers cannot consistently scale lecture delivery across multiple sections while tracking student engagement and learning progress. Students learning remotely often receive passive content without course-specific support, and institutions lack a unified view of attendance, engagement, and assessment performance.\r\n\r\nThis project proposes a hybrid course platform where teachers upload course documents and learning objectives, and the system builds a course-specific knowledge base using Retrieval-Augmented Generation (RAG). It generates lecture scripts and produces lecture videos with slides and a talking avatar (lip-sync), delivered via a student application. During playback, the system records attendance and engagement signals and conducts MCQ-based quizzes\/exams with automated grading. A student learning model updates mastery and confidence to support explainable recommendations and targeted practice.","description":"SmartStudyInstructor is an AI-powered hybrid course delivery platform (web + mobile) that converts teacher-uploaded material into structured lecture delivery with integrated monitoring and assessment. Teachers create a course, define outcomes\/learning objectives, and upload PDFs\/PPTs. The system performs document processing and builds a searchable course knowledge base using RAG (text extraction, chunking, embeddings, vector retrieval). Using retrieved course content, it generates a lecture script and narration. The system then produces a lecture video in which slides appear as the background and an avatar appears on the side with voice narration and lip synchronization, enabling students to consume lectures remotely through the mobile application.\r\n\r\nWhile the lecture video plays, the student application records participation through watch-time and a lightweight, privacy-aware camera-based engagement signal (face presence\/looking-at-screen) sampled periodically during the session. This provides measurable lecture attendance and engagement indicators without storing raw video. Students can ask questions during or after the lecture via a course-specific Q\/A feature; the system answers using retrieved material only and can show citations\/snippets to reduce hallucinations.\r\n\r\nAfter each lecture, the platform conducts MCQ-based quizzes\/exams, auto-grades results, and updates a student learning model (topic mastery, confidence, pace, and engagement trend). Teachers monitor attendance, engagement, performance, and explainable recommendations through a dashboard, allowing them to manage multiple sections with reduced repetitive workload. Optional extensions (OCR checking, plagiarism) are documented as future enhancements, while evaluation focuses on a working, demonstrable hybrid pipeline.$$\nModule 1: Authentication & Role Management\r\nProvides secure login\/logout and role-based access for Teacher and Student (and optional Admin). Controls access to course content, lecture generation, monitoring, and assessments.\r\n\r\nModule 2: Knowledge Intelligence (RAG-Based) Module\r\nPerforms document ingestion (PDF\/PPT), text cleaning, chunking, embeddings generation, and vector database indexing. Supports retrieval of top-k relevant content for lecture generation and Q\/A, with optional citations\/snippet references.\r\n\r\nModule 3: Teacher Portal & Course Management\r\nAllows teachers to create courses\/sections, enroll students, define learning objectives (OBE), upload material, generate\/publish lectures and quizzes, and monitor class analytics.\r\n\r\nModule 4: Lecture Script + Slide Generation Module\r\nGenerates an objective-aligned lecture script and narration script. Produces slides automatically (titles\/bullets) and timestamps for synchronization with narration.\r\n\r\nModule 5: Video Generation Module (Slides + Avatar + Lip-Sync)\r\nGenerates narration audio using free TTS and produces the lecture MP4 by combining slide background with a side avatar panel. Lip-sync is applied to the avatar using a lightweight pipeline; a fallback mode (static avatar + subtitles) ensures video generation reliability.\r\n\r\nModule 6: Student App (Hybrid Lecture Delivery + Q\/A)\r\nDisplays scheduled\/past lectures, streams videos, supports resume playback, and allows course-specific Q\/A during lecture.\r\n\r\nModule 7: Attendance & Engagement Monitoring Module\r\nComputes attendance using watch-time thresholds and periodically sampled camera-based engagement signals. Stores only numeric metrics (no raw video) and supports teacher override.\r\n\r\nModule 8: Assessment & Student Modeling + Analytics\/Explainability\r\nGenerates MCQ quizzes\/exams aligned to objectives, auto-grades results, updates student mastery\/confidence, and presents dashboards for teacher\/student. Maintains explainability logs for recommendations (e.g., weak topic flagged, tier selection).$$\nI will develop the student-facing application and user experience across hybrid learning workflows. This includes authentication integration, the Flutter lecture player with resume and progress tracking, and the live Q\/A interface for course-specific tutoring. I will implement the lecture video pipeline (slide generation, free TTS voiceover, and avatar lip-sync composition with a reliable fallback). Additionally, I will contribute to analytics presentation and explainability visualization in both teacher and student views to ensure decision transparency. My responsibilities focus on usability, smooth lecture delivery, and ensuring the end-to-end hybrid flow is demonstrable with real data.$$\nI will implement the backend architecture and core AI pipeline. This includes RAG knowledge base creation (extraction, chunking, embeddings, vector retrieval) and lecture\/narration generation from retrieved content. I will also implement assessment screens (quiz\/exam attempt flow), results visualization, and student dashboard views (mastery, confidence trend, recommendations). I will also implement attendance and engagement signal processing, MCQ generation and auto-grading, student model updates, and teacher dashboard APIs. My focus will be on demonstrable technical contribution, stable APIs, database integration, and end-to-end system integration and deployment.$$\n$$\n$$\n$$\n$$\n","comments":"","isDraft":1,"status":1,"created_at":"2026-03-04 03:02:04","updated_at":"2026-03-25 23:23:05","isReviewDraft":1,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1062,"project_id":1448,"title":"DevSynx: AI-ERP for software houses with team management, client portal, real-time tracking, and automated workflow.","prob":"Software houses struggle with project management due to poor task assignment, lack of team coordination, and inaccurate timeline estimation, leading to delays, budget overruns, and client distrust. Managers cannot effectively track workloads across teams, while clients remain unaware of project status. Without automation, ensuring task completion is challenging, and lacking workload prediction leads to developer burnout.\r\nDevSynx addresses these issues with an AI-enabled ERP system where managers create teams and assign projects to teams, then team head break projects into tasks and assign it to developers . The Agentic AI Workflow Manager automatically verifies completed work. A Client Dashboard provides real-time progress visibility, milestones, and completion estimates. The AI Workload & Risk Predictor analyzes developer capacity, task complexity, and assignments to predict overloads and risks.","description":"The proposed system is a web-based Enterprise Resource Planning (ERP) solution designed specifically for software houses to efficiently manage software development projects through a structured team-based approach while improving client transparency. The system provides role-based access control, offering separate dashboards and permissions for managers, team heads, developers, and clients. A key feature is the Teams Management , where managers can create multiple development teams, assign team heads, and allocate projects to specific teams. Team heads then break down projects into granular tasks and assign them to individual team members, ensuring clear accountability and structured workflow. The Agentic AI Workflow Manager serves as an automated supervisor that continuously monitors task completion. When a developer marks a task as complete and uploads their work, the AI agent verifies the submission against the task requirements, ensuring that the work has been genuinely completed. It flags any discrepancies for team head or manager review. The AI Workload & Risk Predictor analyzes developer capacity, current workload, task complexity, and historical performance data to predict potential developer overload and associated project risks. It provides intelligent recommendations for task reassignment or timeline adjustments, helping managers maintain balanced workloads and prevent burnout. A dedicated Client Portal Module allows clients to securely log in and view real-time project progress, completed and pending tasks, estimated completion time, and summaries ensuring transparency while keeping internal development details protected progress tracking.$$\nModule 1: Authentication & User Management Module\r\nProvides secure login, logout, and role-based access for managers, team heads, developers, and clients with distinct permissions and dashboard views.\r\nModule 2: Teams Management Module\r\nHave manager portal and to create projects and teams, assign projects to team. Team heads can manage their team members.\r\nModule 3: Project Task Management Module\r\nEnables team heads to break down projects into tasks, assign tasks to specific team members, set deadlines, and track task status updates.\r\nModule 4: Real-Time Progress Tracking Module\r\nAutomatically tracks project and task progress based on live updates from developers, providing real-time visibility into project health and completion status.\r\nModule 5: Client Portal Module\r\nEnables clients to view real-time project progress, completed and pending tasks, estimated delivery time, and high-level project summaries.\r\nModule 6: Agentic AI Workflow Manager & Workload Predictor Module\r\nProvides an AI agent that verifies developer task submissions for completion and compliance, it also has an AI-powered workload predictor that analyzes developer capacity and predicts potential overloads and project risks.$$\nI shall develop the Teams Management Module, Real-Time Progress Tracking Module, and the AI Workload & Risk Predictor. This includes team creation, team head assignment, project allocation to teams, real-time tracking of tasks and projects across all teams, and the AI-powered engine that analyzes developer workload, task complexity, and historical data to predict potential overloads and associated project risks.$$\nI shall develop the Authentication & User Management Module, Project Task Management Module, Client Portal Module, and the Agentic AI Workflow Manager. This includes secure role-based login, project breakdown into tasks, task assignment and status management, the client dashboard for real-time project viewing, and the AI agent that verifies developer uploads against assigned tasks to ensure completion and compliance.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-10 20:05:12","updated_at":"2026-03-10 20:05:12","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1068,"project_id":1458,"title":"AI-Based Intelligent Resume Screening and Job Matching System","prob":"Recruitment is a complex and time-consuming process where recruiters must manually analyze large numbers of resumes to identify suitable candidates. Traditional Applicant Tracking Systems (ATS) rely mainly on keyword matching and lack intelligent decision-making capabilities and transparency in hiring decisions.\r\nThis project proposes an AI-based Intelligent Resume Screening and Job Matching System that automates candidate evaluation by analyzing resume features, computing resume-job matching scores, predicting shortlisting probability, and providing explainable insights into AI decisions.\r\nThe system focuses specifically on screening candidates for Software Engineering and IT-related job roles, improving hiring efficiency, reducing manual workload, and enhancing transparency using Machine Learning, Natural Language Processing (NLP), and Explainable AI techniques.","description":"The AI-Based Intelligent Resume Screening and Job Matching System aims to automate the recruitment screening process for Software Engineering and IT-related job roles using Artificial Intelligence (AI), Machine Learning (ML), and Natural Language Processing (NLP). The system reduces the manual effort required by recruiters by automatically analyzing candidate resumes, matching them with job requirements, and predicting whether a candidate should be shortlisted.\r\nThe system begins when recruiters upload job descriptions and candidate resumes in formats such as PDF, DOC, or DOCX. A resume parsing module extracts relevant information including candidate skills, education, work experience, and projects. This information is then converted into structured features representing candidate qualifications.\r\nThe project uses a hybrid engineered recruitment dataset containing approximately 12,000 candidate samples focused on IT and software engineering roles. The dataset includes features such as semantic similarity score between resume and job description, skills match score, years of experience, education level, project count, GitHub activity score, and a shortlisting label. This dataset is used to train and evaluate the machine learning model.\r\nTo determine how well a resume matches a job description, the system uses Sentence-BERT (SBERT) to generate text embeddings and compute cosine similarity between the resume and job description. These similarity scores, along with other candidate features, are used as input to a Random Forest classification model, which predicts whether a candidate should be shortlisted.\r\nTo improve transparency, the system integrates Explainable Artificial Intelligence (XAI) using SHAP, which identifies how different features influence the model\u2019s decision. The results, including candidate rankings, predictions, and explanation insights, are displayed through an interactive recruiter dashboard that supports efficient and transparent recruitment decision-making.$$\nModule 1 \u2013 Resume Upload and Parsing\r\n* Accept resumes in PDF \/ DOC \/ DOCX formats\r\n* Extract resume text\r\n* Identify candidate skills, education, and experience\r\n\r\nModule 2 \u2013 Feature Engineering and Scoring\r\n* Compute semantic similarity score using NLP embeddings\r\n* Calculate skills match score\r\n* Calculate experience relevance score\r\n* Generate overall AI suitability score\r\n\r\nModule 3 \u2013 Candidate Classification\r\n* Using Random Forest to Predict Shortlisted \/ Not Shortlisted\r\n* Provide probability-based decision\r\n\r\nModule 4 \u2013 Candidate Ranking\r\n* Rank candidates based on AI score\r\n* Recommend top candidates for specific job roles\r\n\r\nModule 5 \u2013 Explainable AI Module\r\n* Apply SHAP on trained model\r\n* Show contribution of skills, experience, projects, and similarity score\r\n\r\nModule 6 \u2013 Recruiter Dashboard\r\n* Upload resumes and job descriptions\r\n* View AI predictions and rankings\r\n* Visualize explainable insights$$\nIn this project I will develop Module 1 \u2013 Resume Upload and Parsing, Module 2 \u2013 Feature Engineering and Scoring, and Module 3 \u2013 Candidate Classification. \r\n\r\nThe Resume Upload and Parsing module will accept resumes in PDF, DOC, and DOCX formats, extract resume text, and identify candidate skills, education, and experience. The Feature Engineering and Scoring module will compute semantic similarity between resumes and job descriptions using NLP embeddings, calculate skills match score and experience relevance score, and generate an overall AI suitability score. The Candidate Classification module will use machine learning techniques to predict whether a candidate should be shortlisted or not shortlisted and provide a probability-based decision for candidate evaluation.$$\nIn this project I will develop Module 4 \u2013 Candidate Ranking, Module 5 \u2013 Explainable AI Module, and Module 6 \u2013 Recruiter Dashboard.\r\n\r\nThe Candidate Ranking module will rank candidates based on the calculated AI suitability score and recommend the top candidates for specific job roles. The Explainable AI module will apply SHAP on the trained model to provide interpretability and show the contribution of features such as skills, experience, projects, and semantic similarity score in the decision-making process. The Recruiter Dashboard module will provide an interface where recruiters can upload resumes and job descriptions, view AI predictions and candidate rankings, and visualize explainable insights for better recruitment decisions.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-13 15:27:28","updated_at":"2026-03-13 15:27:28","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1055,"project_id":1459,"title":"Flocksy:A Unified Social Media Platform Offering Social Networking for Adult and Safe Learning for Children.","prob":"Most social media platforms are designed for adults and provide the same features to all users, regardless of age. Children who access. these platforms are often exposed to inappropriate content, unsafe interactions, and excessive screen time, creating significant risks for their safety and well-being. Parents and educators face challenges in finding a single platform that enables children to engage safely while supporting learning and creativity. Simultaneously, adults require unrestricted social networking features, including posting, messaging, and media sharing. Existing solutions do not adequately address the needs of both age groups within one system. This absence of an age-aware platform results in children being vulnerable online, while adults miss a unified and seamless social experience. There is a pressing need for a platform that provides full social media capabilities for adults alongside a safe, educational, and controlled environment for children, ensuring safety, creativity, and engagement for all users.","description":"Flocksy is a modern social media web application developed using the MERN Stack (MongoDB, Express.js, React.js, Node.js). It introduces a safe, AI-powered, and age-aware experience, connecting users through posts, stories, short videos called Loops, messaging, and following systems, while providing a separate, secure environment for users under 16.\r\n\r\nFor adult users, Flocksy offers full social media functionality. Users can create profiles, upload posts, share stories, like and follow others, view Loops, and send real-time messages. An AI-powered module, FlockMind AI, assists users by suggesting captions, improving content quality, and detecting inappropriate material. This enhances user engagement while ensuring safety and creativity.\r\n\r\nThe standout feature is Kids Mode, automatically activated for users under 16. In this mode, all traditional social media content is hidden, and children access a child-friendly, educational interface. The Kids Home Screen includes large buttons and simple navigation to access stories, quizzes, mini-games, a drawing canvas, and a rewards system. Stories provide moral lessons with text and images only. Quizzes test basic topics like animals, colors, and numbers. Mini-games encourage logical thinking, while the drawing module fosters creativity. Stars and badges reward achievements without any leaderboard, comparison, or social pressure.\r\n\r\nFlocksy addresses a real-world problem: children\u2019s exposure to inappropriate content on social media and lack of safe, integrated digital platforms. By combining adult social networking with a secure, educational environment for children, Flocksy provides a dual-mode system that promotes safe interaction, creativity, and learning.\r\n\r\nThe backend manages authentication, authorization, data storage, and age-based routing. The frontend dynamically renders interfaces based on user roles, ensuring scalability, usability, and security. With its AI integration and Kids Mode, Flocksy is a responsible and innovative social media solution, making it a strong and practical Final Year Project.$$\n1. User Authentication & Authorization Module\r\nThis module manages user registration, login, logout, and secure authentication. It implements role-based authorization to differentiate between adult users and children. Age verification is performed at signup to activate either Adult Mode or Kids Mode. This ensures secure access and controlled system behavior.\r\n\r\n2. User Profile Management Module\r\nThis module allows adult users to create and manage personal profiles. Users can update profile pictures, bio, and account settings. Children profiles are limited and do not include public visibility or social interaction features.\r\n\r\n3. Posts & Media Sharing Module\r\nThis module enables adult users to upload images and text-based posts. Users can like posts and follow other users. Content is stored securely in the database and displayed in the main feed. This module is disabled for children.\r\n\r\n4. Stories Module\r\nThe Stories module allows adult users to share time-limited content visible for 24 hours. In Kids Mode, this module is replaced with a static moral stories section containing predefined educational stories with text and images only.\r\n\r\n5. Loops (Short Video) Module\r\nThis module provides short video-sharing functionality similar to reels. Adult users can upload, view, and interact with short videos called \u201cLoops.\u201d This feature increases engagement and supports modern content consumption trends.\r\n\r\n6. Messaging & Follow System Module\r\nThis module manages real-time messaging between adult users and maintains the follow\/following relationships. It enables private communication and social networking while ensuring data privacy and security.\r\n\r\n7. AI Assistance & Moderation Module (FlockMind AI)\r\nThis module integrates AI to assist users in content creation by suggesting captions and improving text quality. It also performs basic content moderation to detect inappropriate material and supports safety features, especially for underage users.\r\n\r\n8. Kids Home Screen Module\r\nThis module provides a colorful and child-friendly dashboard for users under 16. It contains large buttons and simple navigation to access educational and creative activities. No social media content is displayed in this module.\r\n\r\n9. Kids Quiz Module\r\nThis module offers multiple-choice quizzes for children on basic topics such as animals, colors, and numbers. \r\n\r\n10. Kids Mini Games Module\r\nThis module includes simple educational games such as Guess the Animal and True\/False. \r\n\r\n11. Kids Drawing Module\r\nThis module provides a drawing canvas where children can draw using a mouse or touch input. A color picker allows creative expression, and drawings may optionally be saved.\r\n\r\n12. Rewards & Badge Module\r\nThis module tracks children\u2019s activities and awards stars for correct quiz answers and game completions. Badges are unlocked after earning a set number of stars. No leaderboards or comparisons are included to avoid pressure.\r\n\r\n13. Admin & Content Management Module\r\nThis module allows administrators to manage users, monitor content, and update kids\u2019 educational material. It ensures system maintenance, moderation, and overall platform control.\r\n14.OTP-Based Secure Mode Switching\r\nTo ensure security:\r\nWhen a kid attempts to switch to Parent Mode\r\nSystem sends OTP to registered parent email\/phone\r\nParent must enter OTP to allow switching\r\nWithout OTP \u2192 access denied$$\nMember 1 \u2013 Adult Features & Core Backend\r\nFocus: Social media features for adults + authentication + AI integration + database management\r\nUser Authentication & Authorization Module\r\n\r\n\r\nRegistration, login\/logout, role-based access\r\n\r\n\r\nAge verification (Kids vs Adult)\r\n\r\n\r\nUser Profile Management Module\r\n\r\n\r\nAdult profile creation and update\r\n\r\n\r\nPosts & Media Sharing Module\r\n\r\n\r\nImage\/text posts, likes, follow\/following\r\n\r\n\r\nStories Module (Adult)\r\n\r\n\r\n24-hour story sharing\r\n\r\n\r\nLoops (Short Video) Module\r\n\r\n\r\nUpload and view short videos\r\n\r\n\r\nMessaging & Follow System Module\r\n\r\n\r\nReal-time messaging\r\n\r\n\r\nFollow\/following management\r\n\r\n\r\nAI Assistance & Moderation Module (FlockMind AI)\r\n\r\n\r\nCaption suggestions, content moderation\r\n\r\n\r\nAdmin & Content Management Module\r\n\r\n\r\nAdmin controls, monitor adult & kids content\r\n\r\n\r\nMember 1 handles: adult features, AI integration, backend logic, database storage, and admin functionality.$$\nMember 2 \u2013 Kids Mode & Frontend \/ UI\r\nFocus: Safe kids environment, interactive modules, front-end UI, and rewards system\r\nKids Home Screen Module\r\n\r\n\r\nChild-friendly dashboard, big buttons, colorful UI\r\n\r\n\r\nKids Quiz Module\r\n\r\n\r\nMCQs, scoring system\r\n\r\n\r\nKids Mini Games Module\r\n\r\n\r\nGuess the Animal, True\/False games\r\n\r\n\r\nKids Drawing Module\r\n\r\n\r\nDigital canvas, color picker, save option\r\n\r\n\r\nRewards & Badge Module\r\n\r\n\r\nStar\/badge system for achievements\r\n\r\n\r\nNo leaderboard\r\n\r\n\r\nFront-end Dynamic Rendering\r\n\r\n\r\nAge-based routing (Kids Mode vs Adult Mode)\r\n\r\n\r\nResponsive design, overflow\/text issues fixes\r\n\r\n\r\nMember 2 handles: all kids modules, UI\/UX, interactive frontend, and rewards logicm$$\n$$\n1.Loom Centric-Socail Media Post Scheduling and Management Platform\r\n\r\n2.Readinook:A social Media platform for Readers$$\n1.Age-Aware Dual Mode System\r\nAutomatically redirects users under 16 to a safe, child-friendly interface while adults access full social features.$$\n2.Interactive Kids Mode\r\n\r\nProvides moral stories, quizzes, mini-games, drawing canvas, and rewards to create a safe, educational, and fun environment for children.$$\n3.AI-Powered Content Assistance (FlockMind AI\r\n\r\nProvides moral stories, quizzes, mini-games, drawing canvas, and rewards to create a safe, educational, and fun environment for children.)","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-04 07:32:18","updated_at":"2026-03-04 07:32:18","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1070,"project_id":1460,"title":"AI-Based Recruitment Platform for Resume Analysis, Real-Time AI Interviews, and Candidate Evaluation.","prob":"Traditional recruitment processes are time-consuming, manual, and often influenced by human bias. Recruiters usually review hundreds of resumes manually, schedule interviews, and evaluate candidates subjectively. This leads to inconsistent hiring decisions, delays in recruitment, and difficulty in identifying the most suitable candidates.\r\n\r\nThis Final Year Project aims to solve these problems by developing an AI-based recruitment system that automates resume screening, conducts AI-assisted interviews, and provides data-driven hiring recommendations. The system will analyze resumes using machine learning techniques and rank candidates based on job requirements.\r\n\r\nAdditionally, the system will conduct automated AI interviews using voice-based interaction. Candidate responses will be converted into text using streaming speech recognition and evaluated using natural language processing techniques. A monitoring module will also detect suspicious behavior such as tab switching and abnormal head movement during interviews. The platform aims to improve hiring efficiency, reduce bias, and help organizations make accurate and transparent recruitment decisions.","description":"The AI Recruitment System is a web-based platform designed to automate and improve the recruitment process using artificial intelligence and machine learning. Employers can create accounts, post job vacancies, and configure evaluation criteria for candidate selection. Candidates can register on the system and apply for available job positions by uploading their resumes.\r\n\r\nOnce a resume is uploaded, the system processes it using natural language processing techniques to extract important information such as skills, education, and experience. The resume is then compared with job requirements using cosine similarity to determine how well the candidate matches the job. Based on this analysis, candidates are automatically ranked and shortlisted.\r\n\r\nShortlisted candidates are invited to participate in an AI-based interview. During the interview, questions are generated using a trained dataset of technical interview questions related to the required skills. The interview is conducted through voice interaction where candidate responses are captured using streaming speech recognition and converted into text for further analysis.\r\n\r\nThe system evaluates candidate answers using natural language processing and scoring techniques. Additionally, a monitoring module runs during the interview to detect suspicious behavior such as multiple tab switching and abnormal head movement using browser events and webcam-based face detection.\r\n\r\nFinally, the system provides recruiters with a decision support dashboard where they can view candidate rankings, resume scores, interview performance, and detailed evaluation reports. The system includes secure login, authentication, and role-based access control to ensure data security and privacy.$$\n(i) User & Role Management\r\nThis module manages authentication and access control in the system. Users can register and log in securely as either recruiters or candidates. Role-based access control ensures that recruiters can manage job postings and candidate evaluations, while candidates can apply for jobs and participate in interviews. The module also manages user profiles, passwords, and session management.\r\n\r\n(ii) Job Posting & Evaluation Configuration\r\nRecruiters can create and manage job postings through this module. They can specify job descriptions, required skills, experience levels, and evaluation criteria. Recruiters can also define weights for resume scoring and interview evaluation factors. This module allows organizations to customize hiring requirements according to their recruitment needs.\r\n\r\n(iii) Resume Processing, AI Shortlisting & Job Recommendation\r\nThis module processes uploaded resumes using natural language processing techniques. It extracts key information such as candidate skills, qualifications, and work experience. The system compares the extracted data with job requirements using cosine similarity to measure resume-job matching. Based on the similarity score, candidates are ranked and automatically shortlisted. The system can also recommend suitable jobs to candidates based on their resume profiles.\r\n\r\n(iv) Real-Time AI Interview, Monitoring & Evaluation\r\nThis module conducts automated interviews for shortlisted candidates. Interview questions are generated using a trained dataset of technical interview questions related to the job skills. The interview is conducted through voice interaction where candidate responses are recorded and processed using streaming speech recognition to generate real-time transcripts.\r\n\r\nThe system evaluates the answers using natural language processing techniques to measure relevance and quality. During the interview, a monitoring module also runs to detect suspicious behavior. It identifies multiple tab switching using browser visibility detection and detects abnormal head movement using webcam-based face detection. If suspicious behavior is detected, the system records it as a potential cheating event.\r\n\r\n(v) Final Evaluation, Decision Support & Reporting\r\nThis module combines resume scores, interview evaluation results, and monitoring reports to produce a final candidate ranking. Recruiters can view detailed evaluation reports including resume similarity scores, interview performance, and behavioral monitoring results. The system also provides explainable AI-based decision support by showing how final scores are calculated, helping recruiters make transparent and informed hiring decisions.$$\nIn the AI-Powered Recruitment System, I shall develop the Resume Processing, AI Shortlisting & Job Recommendation module. This module will allow candidates to upload resumes and will automatically extract important information such as skills, education, and work experience using natural language processing techniques. The system will compare candidate resumes with job requirements using cosine similarity to measure matching scores. Based on these scores, the system will automatically rank and shortlist candidates. Recruiters will be able to review the shortlisted candidates and approve or reject them. Additionally, the system will recommend suitable job opportunities to candidates based on the similarity between their resumes and job descriptions.$$\nIn the AI-Powered Recruitment System, I shall develop the Real-Time AI Interview, Monitoring & Evaluation module. This module will generate interview questions using a dataset of technical interview questions related to job skills. The interview will be conducted using real-time voice interaction where candidate responses are captured and converted into text using streaming speech recognition. The system will evaluate candidate answers using natural language processing techniques to measure correctness and relevance. Additionally, the module will include a monitoring system that detects suspicious behavior such as multiple tab switching and abnormal head movement during interviews. The system will record interview scores, monitoring results, and candidate performance for final evaluation.$$\n$$\n1. Pluto: A Web-Based Job Applicant Management System$$\nReal-time AI-driven voice interview system that conducts live interviews, captures voice responses, generates transcripts, and evaluates technical and communication skills automatically.$$\nRecruiter-controlled dynamic evaluation framework allowing customizable resume criteria, interview factors, and weight distributions for hiring decisions.$$\nExplainable AI-based decision support system that provides transparent score breakdowns for resumes and interviews and justifies final candidate rankings instead of giving basic recommendations.","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-13 19:49:38","updated_at":"2026-03-13 19:49:38","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1059,"project_id":1461,"title":"FINSIGHT\r\nFinancial Intelligence System for Investment & Growth Tracking","prob":"Investors spend countless hours manually researching companies across scattered sources such as financial reports, news, and market prices. They struggle to continuously monitor their portfolios and often lack intelligent guidance for effective decision-making. Most existing platforms only present raw data without meaningful automated analysis, require manual tracking to identify important events, and do not provide any conversational interface for user queries. This results in information overload and delayed responses to rapid market changes.\r\nOur system addresses these limitations by deploying AI agents that automatically analyze company data from multiple sources, continuously monitor portfolios using intelligent alerts, and present real-time visual analytics through interactive dashboards. In addition, a conversational AI interface enables users to ask questions such as \u201cWhy did my stock drop?\u201d and receive instant, data-driven explanations. This approach makes investment analysis and portfolio monitoring more accessible, efficient, and decision-oriented.","description":"This project implements an intelligent multi-agent system for automated investment analysis and continuous portfolio monitoring. Its primary goal is to transform complex financial and market data into clear, actionable insights, enabling users to make informed investment decisions without requiring deep financial expertise.\r\n\r\nThe system consists of five collaborative AI agents:\r\n\r\n1. The Data Collection Agent will employ Selenium-based web scraping to gather financial data directly from web sources without relying on external APIs. It will parse the embedded JSON on Investing.com to extract financial statements such as the income statement, balance sheet, and cash flow statement. In addition, the agent will scrape the KSE-100 table from Sarmaaya.pk, collecting key market metrics including current price, price change, trading volume, and market capitalization. Now the system is fesasible and data collection issue is ressolved.\r\n\r\n2. The Analyst Agent uses Large Language Models to interpret financial statements, historical performance, and daily market data to evaluate company health, valuation, and risk, producing concise investment summaries highlighting strengths, risks, and trends.\r\n\r\n3. The Monitoring Agent continuously tracks user-selected companies and portfolios, detecting significant movements such as abnormal volatility or valuation shifts and generating intelligent alerts.\r\n\r\n4. The News Aggregation Agent collects and summarizes relevant financial news, linking market movements to real-world events for better context.\r\n\r\n5. The Conversational AI Agent provides a natural language interface, allowing users to ask questions like \u201cWhy did this stock rise today?\u201d and receive data-backed explanations.\r\n\r\nBy combining annual fundamental data with daily market behavior, the system delivers a comprehensive investment view through dashboards, alerts, and conversational responses, making investment analysis faster, more accessible, and intelligent.$$\nModule 1: Real-Time Data Collection & Processing Pipeline\r\n\r\nMulti-source data collection agents using web scraping\r\nPSX-focused financial statement scraping using Selenium\r\nNews aggregation from multiple sources, primarily via web scraping\r\nData validation and cleaning pipelines for consistency and quality checks\r\nETL pipeline to transform and load structured data into storage systems\r\n\r\n\r\n\r\nModule 2: Multi-Agent Analysis System\r\n\r\nResearch Agent that orchestrates data collection and analysis workflows\r\nLLM integration (OpenAI\/Claude API) for generating natural language insights\r\nAutomated company analysis from financial statements(Balance Sheets, Income statements and cash flow statements), news, and market data\r\nMulti-company comparative analysis engine\r\nSector and industry trend identification\r\nNatural language report generation\r\nAgent coordination and task scheduling system\r\n\r\n\r\n\r\nModule 3: Continuous Monitoring & Alert System\r\n\r\nReal-time portfolio tracking \r\nIntelligent alert engine for price movements, news events, volume spikes\r\nPerformance metrics calculation (returns, gains\/losses, allocation)\r\nDaily\/weekly automated report generation using LLM\r\nEvent detection system (earnings announcements, news mentions)\r\nMulti-condition monitoring (price thresholds, percentage changes)\r\nAlert delivery system (in-app notifications, email)\r\nHistorical tracking and trend analysis\r\n\r\n\r\n\r\nModule 4: Conversational AI & Natural Language Interface\r\n\r\nLLM-powered chatbot for fact-based investment queries\r\nIntent classification for accurate data routing\r\nContext-aware conversations using verified data\r\nNatural language understanding of financial questions\r\nReal-time data retrieval and synthesis\r\nEvidence-based explanations backed by financial metrics\r\nMulti-turn conversations with data consistency\r\nPersonalized guidance using portfolio data\r\nChat history and session management for traceability\r\n\r\n\r\n\r\nModule 5: Frontend Interface & Visualization System\r\n\r\nUnified React.js frontend for the complete investment analysis platform\r\nUser authentication and portfolio management interface\r\nInteractive dashboards for portfolio performance and company analysis\r\nData visualizations for trends, comparisons, and holdings breakdown\r\nReal-time alerts and monitoring feed\r\nIntegrated chatbot interface for data-driven user queries\r\nMarket news and event visualization\r\nDrill-down views for detailed analysis and insights\r\nResponsive design for desktop and mobile devices\r\n\r\n\r\n\r\nModule 6: API Layer & System Integration\r\n\r\nBackend API layer developed using FastAPI or Django.\r\nAPIs to expose financial data, analysis results, and user portfolios\r\nCentralized API layer connecting frontend, AI agents, and data sources\r\nAsynchronous request handling for efficient data processing\r\nSecure user authentication and authorization mechanism\r\nInter-module communication for agent coordination and task execution\r\nBasic caching mechanism to improve response time\r\nError handling and logging for system reliability and debugging\r\nAPI documentation to support development and integration\r\nSystem monitoring and health checks$$\n(Modules 2, 4, and 6)\r\nI will design and implement the core intelligence and backend architecture of the system. This includes developing the multi-agent analysis framework that coordinates analysis workflows, performs automated company and sector analysis, and generates structured insights using large language models. I will also develop the conversational AI interface to support natural language queries, intent understanding, multi-turn context handling, and personalized responses. Additionally, I will implement the API layer and system integration, exposing backend functionalities through secure REST APIs and managing agent coordination, logging, error handling, and system reliability.$$\n(Modules 1, 3, and 5)\r\nI will develop the real-time data collection and processing pipeline, including multi-source data ingestion, validation, cleaning, scheduling, and structured storage of financial and time-series data. I will also implement the continuous monitoring and alert system to track portfolio performance, detect significant market changes, and generate automated alerts with historical trend analysis. In addition, I will design and build the interactive dashboard and visualization system, providing real-time insights, charts, comparison views, responsive user interfaces, and portfolio management features.$$\n$$\nStock Market Analyzer$$\nSimilar FYP tries to predict future prices using ML models that often fail when markets change, while our project analyzes real data and explains it in simple language, always reliable and useful.$$\nSimilar FYP needs to build and train separate models for every single company, while our project uses smart agentic system that automatically handles many companies without any extra setup.$$\nSimilar FYP shows charts and numbers on a screen and leaves you to figure out what they mean, while our project lets you ask questions like \"why is my portfolio down?\" and gives you clear answers.","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-09 05:41:00","updated_at":"2026-03-10 08:44:52","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1072,"project_id":1435,"title":"Smart AI-Assisted Team Task Management Platform for Software Houses with Progress Tracking and Reporting","prob":"Many software development teams face difficulties managing daily tasks and tracking project progress. In many organizations, tasks are communicated through emails, messages, or spreadsheets, which often creates confusion, missed deadlines, and lack of accountability. Team members may not clearly understand their assigned work, task priorities, or project deadlines, while managers find it difficult to monitor real-time progress of tasks and team members. This project proposes a centralized web-based task management platform for software development teams. The system allows managers to create tasks, assign them to members, set priorities, and define due dates. Team members can view tasks, update progress using a checklist, and mark tasks as completed. The system automatically tracks task status and generates progress reports. An AI chatbot assistant also helps users understand task requirements, clarify instructions, and provide work-related guidance, while the dashboard shows visual insights about task distribution and completion status.","description":"The proposed system is a web-based Task Management Platform designed for software development teams to organize work, assign tasks, and monitor project progress in a structured way. Users can register and securely log in to the system. After logging in, administrators or project managers can create tasks and assign them to one or multiple team members. Each task contains details such as title, description, priority level, due date, and a checklist that divides the task into smaller steps. Team members can access their personal dashboard to view assigned tasks and monitor their progress. As checklist items are completed, the system automatically calculates task progress and updates the task status. Tasks move through stages such as pending, in-progress, and completed, allowing managers to easily track team performance and project progress. The platform also provides analytical dashboards with visual charts that display task distribution, priority levels, and completion statistics to help managers identify delayed tasks. Additionally, the system supports file attachments where users can add links to documents or resources related to each task so that all task information remains organized in one place. A key feature of the platform is an AI chatbot assistant that helps users understand task descriptions, clarify instructions, and provide guidance related to their assigned work. Another important feature is the report generation module that allows administrators to download task reports for monitoring and evaluation. The system will be developed using the MERN stack with React for the user interface, Node.js and Express for backend APIs, and MongoDB for storing application data.$$\n1. Authentication and User Management Module:\r\nThis module manages user registration, login, and profile management. Users can create accounts and securely access the platform using authentication APIs. The module also allows administrators to view all registered users and manage user information such as profile images and user details.\r\n\r\n2. Dashboard and Analytics Module:\r\nThe dashboard provides an overview of system activity and task progress. It displays important statistics such as total tasks, completed tasks, pending tasks, and task priorities. Visual charts such as pie charts and bar charts help managers understand work distribution and monitor project performance.\r\n\r\n3. Task Management Module:\r\nThis module allows administrators or project managers to create tasks with detailed information such as title, description, priority level, and due date. Tasks can be assigned to one or multiple users. The system also allows updating, deleting, and modifying tasks when project requirements change.\r\n\r\n4. Task Tracking and Checklist Module:\r\nEach task includes a checklist that divides the task into smaller steps. Team members can mark checklist items as completed while working on the task. Based on checklist completion, the system automatically calculates task progress and updates the task status. This helps managers monitor the progress of tasks in real time.\r\n\r\n5. Team Collaboration Module:\r\nThe system supports assigning tasks to multiple users so that team members can collaborate on shared work. Each user can view tasks assigned to them through their personal dashboard and update their progress accordingly.\r\n\r\n6. File Attachment Module:\r\nUsers can attach file links related to tasks such as design documents, requirement specifications, or other reference materials. This ensures that all important information related to a task is stored and accessible in one place.\r\n\r\n7. Report Generation Module:\r\nThis module allows administrators to generate and download task reports. These reports help managers evaluate team productivity, monitor project performance, and maintain documentation for future reference.\r\n\r\n8. Responsive User Interface Module:\r\nThe system will include a clean and mobile-responsive interface developed using React and Tailwind CSS. The interface will contain a sidebar navigation menu, dashboards, and task cards to make navigation easy and user-friendly across different devices.\r\n\r\n9. AI Chatbot Assistance Module:\r\nThis module integrates an AI-powered chatbot into the platform to assist users while working on their tasks. The chatbot helps team members understand task descriptions, clarify project requirements, and provide suggestions related to their assigned work. Users can ask questions about their tasks, and the chatbot provides helpful explanations and guidance. This feature reduces confusion, improves communication, and helps team members complete tasks more efficiently.$$\nHaider Ali Moeen will develop the backend system using Node.js and Express. His responsibilities include designing MongoDB database schemas for users and tasks and creating REST APIs for authentication, task management, and user management. He will implement APIs for creating, updating, deleting, and retrieving tasks from the database. He will also develop the report generation functionality and implement backend logic for task tracking and checklist updates while ensuring secure data handling.$$\nMuhammad Muzammil Tufail will develop the frontend interface using React and Tailwind CSS. His responsibilities include designing the login and signup pages, implementing the dashboard layout, and creating task management pages. He will build the task creation form, task cards, and task update interfaces. He will also integrate frontend components with backend APIs using Axios to fetch and display real-time data. Additionally, he will implement dashboard charts to visualize task progress, priority distribution, and recent task activity.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-13 22:12:01","updated_at":"2026-03-13 22:12:01","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1066,"project_id":1457,"title":"Tez rider","prob":"Pakistan\u2019s e-commerce sector faces major issues of trust, accountability, and delivery delays, including unverified riders, late deliveries even within nearby areas, lack of real-time tracking, damaged or incorrect orders without proof, and weak dispute resolution. Overseas Pakistanis face difficulty in sending essentials with confidence. Tez rider solves these gaps through a trust-focused delivery ecosystem that assigns orders to the nearest verified rider instead of lengthy courier processes, ensuring faster delivery. OTP confirmation, masked CNIC verification, GPS tracking, digital proof of delivery, checklists, and structured feedback ensure transparency, reliability, and empowerment of the local workforce. Providing secure job opportunities for riders.","description":"Tez rider is a verified delivery platform designed to solve Pakistan\u2019s e-commerce trust issues through multi-layered security and accountability. It supports four roles\u2014customers, vendors, riders, and administrators\u2014across groceries, medicines, gifts, and daily essentials. Orders are assigned to nearby verified riders with real-time GPS tracking and ETA visibility. Customers can view rider identity (photo, rating) before delivery. Mandatory checklists, OTP-based confirmation, and digital proof of delivery (time, location, photo) ensure transparency. Integrated payments, structured feedback, and formal dispute resolution complete an end-to-end trusted delivery ecosystem.$$\nTez rider Modules includes:\r\n(i) User Management & Authentication System: Handles registration, login, and role-based access control for all four user types (Customer, Vendor, Rider, Admin) with JWT token-based authentication and profile management.\r\n(ii) Order Management & Processing System: Manages the complete order lifecycle including product browsing, cart management, order placement, status tracking, and order history with multi-category support.It also supports Multi-Rider Collaborative Delivery (MRCD), where multi-vendor orders are split by location, vendors are grouped using K-Means clustering, and nearby riders are assigned. The order is divided into smaller tasks, and each rider is assigned a specific pickup or delivery part. This allows tasks to be completed in parallel, improving efficiency and reducing overall delivery time. Customers can track multiple riders, with a single OTP for final delivery confirmation.\r\n(iii) Real-Time Tracking & Delivery Verification System: Provides live GPS tracking of riders, displays rider identity , implements OTP-based delivery confirmation, and captures digital proof of delivery with photo, timestamp, and location.If a rider goes offline, the system saves updates locally and shows the last known location. When the internet reconnects, the data is synchronized with the server. If the rider stays offline too long, the system uses the Haversine algorithm to find the nearest rider and a Greedy approach to quickly reassign the order. Calculates distance using the Haversine algorithm and updates ETA dynamically based on speed and traffic.\r\n(iv) Vendor Management & Product System: Enables vendor registration with business verification, product inventory CRUD operations, order acceptance\/rejection, and performance analytics with ratings.\r\n(v) Rider Management & Assignment System: Handles rider registration, implements intelligent location-based assignment algorithm, manages availability status, and tracks delivery performance and earnings.Nearby riders within the city (e.g., Islamabad) are assigned based on proximity and availability to ensure faster delivery.\r\n(vi) Payment Integration & Transaction Management: Integrates multiple payment gateways, processes secure transactions, generates digital invoices, and manages refunds for cancelled or disputed orders.\r\n(vii) Feedback, Rating & Dispute Resolution System: Implements dual feedback mechanism with public ratings and private complaints, manages dispute resolution workflows, and enables blocking of low-rated vendors or riders.\r\n(viii) Admin Dashboard & Analytics System: Provides comprehensive system oversight with approval workflows, real-time order monitoring, analytics dashboards with charts, and business intelligence reports showing performance metrics.A separate dashboard will allow riders and vendors to register. Riders must provide CNIC, driving license, vehicle registration, contact info, and bank details. Vendors must submit CNIC, business license, shop info, contact info, bank details, and product categories. The admin manually verifies all documents before approval, which enhances security and ensures only verified users can access the system.\r\n(ix) Notification & Communication System: Delivers multi-channel notifications through Firebase Cloud Messaging for push notifications, SMS gateway for OTP delivery, and email for order confirmations and invoices.\r\n(x) Security & Data Protection System: Implements end-to-end encryption, CNIC masking for privacy, secure file storage, input validation, SQL injection prevention, and audit logging for critical operations.$$\nIn Tez rider, I shall develop User Management & Authentication System, Order Management & Processing System, and Real-Time Tracking & Delivery Verification System which includes: (i) user registration for all four roles (Customer, Vendor, Rider, Admin) with JWT authentication, (ii) role-based access control and profile management, (iii) product catalog with search and filtering, (iv) shopping cart and order placement, (v) intelligent rider assignment algorithm, (vi) live GPS tracking with Google Maps API, (vii) OTP-based delivery confirmation, (viii) digital proof of delivery capture.$$\nIn Tez rider, I shall develop Vendor\/Rider Management, Payment Integration, Admin Dashboard, and Security Systems which includes: (i) vendor registration with business verification and product inventory CRUD operations, (ii) rider registration and location-based assignment, (iii) payment gateway integration with transaction processing, (iv) dual feedback mechanism with ratings and complaints, (v) admin panel with analytics dashboards and charts, (vi) Firebase Cloud Messaging and SMS gateway for notifications, (vii) data encryption and security measures.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-11 23:32:21","updated_at":"2026-03-11 23:32:21","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1077,"project_id":1462,"title":"Doctor AI: An Intelligent AI-Driven Web Platform for Mental Health Support and Emotional Analysis","prob":"Many individuals in Pakistan, particularly in urban areas like Islamabad, face significant mental health challenges such as anxiety and depression but lack affordable, immediate, or confidential support. Traditional healthcare systems focus on physical ailments, leaving a critical gap in emotional well-being resources. Barriers include social stigma, high consultation costs, limited availability of licensed professionals, and absence of accessible digital mental health tools. Doctor AI addresses this by providing 24\/7 AI-powered mental health support through a fine-tuned chatbot, a real-time sentiment and anxiety detection engine that adapts responses to the user's emotional state, a GAD-7 based anxiety tracker, a crisis detection layer that surfaces emergency helplines when high-risk language is detected, and a localized directory of verified psychologists in Islamabad, helping users transition from self-guided support to professional clinical treatment when necessary.","description":"Doctor AI is a full-stack web-based mental health platform providing empathetic, AI-powered emotional support through seven integrated modules. The system's core is an intelligent chatbot built on a fine-tuned T5 Transformer model, trained on the Counsel Chat and EmpatheticDialogues datasets, complemented by the Groq API (LLaMA 3.3 70B) as a fallback for open-ended dialogue. A real-time Sentiment and Anxiety Detection Engine runs locally using a HuggingFace distilroberta emotion classifier and NLTK VADER, with a dedicated binary anxiety classifier and a digitized GAD-7 questionnaire for clinically grounded anxiety tracking. A Recommendation Engine uses spaCy NER to extract mental health terms from user messages and delivers personalized self-help resources adapted to the user's detected emotional state. An Adaptive UI built in React and Next.js dynamically changes themes, renders emotion-specific content cards, and includes crisis detection that surfaces emergency helpline information when high-risk language is detected. An automated web scraping pipeline using Scrapy and Beautiful Soup maintains an up-to-date directory of verified psychologists in Islamabad stored in MongoDB. A User Analytics Dashboard visualizes emotional trends and GAD-7 scores over time using Plotly. An Admin Dashboard provides platform oversight, professional listing approval, resource library management, and anonymized platform-wide sentiment analytics. Training data is sourced entirely from established public datasets on HuggingFace and Kaggle, no scraping is required for model training. The backend is built with FastAPI and Node.js, with MongoDB Atlas as the database. The entire project operates at zero cost using free and open-source tools.$$\nThe project comprises seven modules:\r\n1.\tAI Conversational Engine: A T5-small model fine-tuned using the HuggingFace Trainer API on Counsel Chat and EmpatheticDialogues datasets for empathetic mental health dialogue. The Groq API (LLaMA 3.3 70B) serves as a fallback layer triggered when the fine-tuned model confidence is below threshold. Multi-turn context is stored in MongoDB.\r\n2.\tSentiment and Anxiety Detection Engine: A local two-model ensemble (HuggingFace distilroberta emotion classifier and NLTK VADER) processes every user message with no API costs. A dedicated binary anxiety classifier trained on the Mental Health Reddit Dataset and GoEmotions detects anxiety-specific language. A GAD-7 questionnaire provides periodic clinical scoring. Crisis keywords trigger automatic escalation to emergency helplines.\r\n3.\tRecommendation Engine: spaCy NER extracts medical terms from user input. A sentiment-driven mapping delivers top personalized self-help resources from a MongoDB library tagged by emotion. Resources adapt as user sentiment changes across the session.\r\n4.\tAdaptive UI and Content Delivery: Built in React and Next.js with Tailwind CSS. Dynamically switches themes and renders content cards based on real-time sentiment scores. Includes JWT-based authentication and crisis detection UI.\r\n5.\tLocalized Resource Scraper and Directory: Scrapy and Beautiful Soup periodically crawl public hospital and psychology association websites in Islamabad. Verified psychologist records are stored and deduplicated in MongoDB.\r\n6.\tUser Analytics Dashboard: MongoDB Aggregation computes time-series emotion trends. Plotly renders emotion frequency charts, mood trend lines, and GAD-7 score history with PDF export.\r\n7.\tAdmin Management Dashboard: Role-based JWT authentication. Admins manage users, approve scraped professional listings, manage the self-help resource library, and monitor anonymized platform-wide sentiment trends.$$\nI will develop the frontend and user experience layer. This includes: designing the full React and Next.js interface with Tailwind CSS for the chatbot, assessment forms, and directory pages; implementing JWT-based login and signup; building the Adaptive UI logic for sentiment-driven theme switching and content card rendering; developing the Recommendation Engine frontend for resource display; building the User Analytics Dashboard with Plotly and Chart.js including PDF export; and designing the Admin Dashboard portal for user management, directory review, and content management.$$\nI will develop the AI core and backend. This includes: fine-tuning the T5 model on Counsel Chat and EmpatheticDialogues using the HuggingFace Trainer API on Google Colab and deploying via FastAPI; building the sentiment and anxiety detection ensemble using distilroberta, NLTK VADER, and a binary anxiety classifier with GAD-7 scoring and crisis escalation logic; integrating the Groq API with prompt engineering and fallback logic; building the Recommendation Engine backend with spaCy NER and sentiment-to-resource mapping; developing the Admin Dashboard backend with role-based auth and aggregation queries; and building the Scrapy scraping pipeline with full MongoDB schema and FastAPI routes.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-14 20:01:27","updated_at":"2026-03-14 23:15:25","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1076,"project_id":1463,"title":"AutoGrade","prob":"Manual evaluation of handwritten examination papers is a time-consuming, error-prone, and inconsistent process in educational institutions. Teachers often spend significant effort checking answers, assigning partial marks, calculating grades, GPA, class averages, and preparing CLO-based performance reports. This workload increases with large class sizes and tight academic schedules, leading to delayed result publication and possible human bias or calculation mistakes. Additionally, institutions lack automated tools to analyze overall class performance, grade distribution, question-wise difficulty, and CLO achievement, which are essential for outcome-based education and curriculum improvement. Existing systems mainly focus on result storage and fail to provide intelligent evaluation and academic analytics. This FYP addresses these challenges by automating the grading of handwritten exams and generating detailed academic insights such as GPA calculation, grade statistics, question-wise analysis, and CLO-based performance evaluation, thereby improving accuracy, efficiency, transparency, and decision-making in the examination process.","description":"This project proposes an AI-based system for automated evaluation and analysis of handwritten examination papers, assignments, and quizzes. The system aims to reduce the manual workload of instructors while improving accuracy, consistency, and transparency in academic assessment. Teachers can upload scanned answer sheets, which are processed to extract both objective and subjective responses.\r\nThe system automatically evaluates MCQ-based questions using predefined answer keys and assesses descriptive answers through semantic analysis and rubric-based scoring, with support for partial marking. A human-in-the-loop mechanism allows instructors to review, adjust, and approve marks before finalization.\r\nBeyond grading, the system provides comprehensive academic analytics. It calculates grades, GPA, class average, maximum, minimum, as well as best, average, and worst performance for assignments, quizzes, and examinations. Grade distribution, question-wise performance analysis, and identification of difficult questions are also supported. Additionally, the system enables CLO-based performance evaluation by mapping questions to learning outcomes and measuring class achievement for each CLO.\r\nStudents can view their results, feedback, and performance summaries through a dedicated dashboard, while instructors gain actionable insights to improve teaching strategies and curriculum planning. Overall, the system enhances efficiency, reduces human error, and supports outcome-based education through intelligent grading and performance analysis.$$\n1) User & Role Management Module: This module manages secure login and authentication for teachers and students.(i) role-based access control, (ii) secure login and session handling, (iii) separate dashboards for teachers and students, and (iv) basic profile and account management.\r\n2) Answer Sheet Upload & Management Module: This module manages scanned answer sheet submissions.(i) upload support for image and PDF formats, (ii) batch upload for multiple submissions, (iii) linking answer sheets with student records and exams, and (iv) organized storage for easy retrieval.\r\n3) Image Preprocessing Module: This module improves scanned answer sheet quality before text extraction.(i) noise removal from images, (ii) grayscale conversion, (iii) skew correction for tilted pages, and (iv) contrast enhancement to improve text clarity.\r\n4) Optical Character Recognition (OCR) Module: This module converts handwritten and printed text into digital form.(i) text detection from scanned papers, (ii) segmentation of answers by question, (iii) conversion to machine-readable text, and (iv) structured storage for evaluation.\r\n5) MCQ Detection & Auto-Checking Module: This module automatically grades MCQ answers.(i) detection of marked options, (ii) comparison with answer keys, (iii) automatic mark assignment, and (iv) quick calculation of MCQ scores.\r\n6) Descriptive Answer Evaluation Module: This module evaluates subjective answers using AI techniques. (i) semantic similarity comparison with model answers, (ii) analysis of concept coverage and relevance, (iii) assessment of completeness, and (iv) intelligent scoring of responses.\r\n7) Rubric-Based Grading Module: This module applies structured grading rubrics.(i) teacher-defined grading criteria, (ii) partial marking support, (iii) consistent scoring across students, and (iv) flexible grading rules.\r\n8) Cheating & Answer Similarity Detection Module: This module identifies possible answer copying.(i) similarity comparison between answers, (ii) detection of highly similar responses, (iii) automatic flagging of suspicious cases, and (iv) reporting for teacher review.\r\n9) Automated Feedback Generation Module: This module generates feedback for students. (i) question-wise feedback generation, (ii) identification of missing concepts, (iii) improvement suggestions, and (iv) delivery of feedback through the student dashboard.\r\n10) Teacher Review & Human-in-the-Loop Module: This module allows teachers to verify AI grading. (i) viewing evaluated answers, (ii) confidence score display, (iii) mark adjustment and comments, and (iv) final approval before publishing results.\r\n11) Rechecking & Query Management Module: This module manages student rechecking requests.(i) submission of rechecking requests, (ii) teacher reassessment of answers, (iii) mark updates if needed, and (iv) transparent record of changes.\r\n12) Analytics & Performance Reporting Module: This module provides academic performance insights. (i) class statistics such as average and highest marks, (ii) grade distribution reports, (iii) question-wise analysis, and (iv) identification of weak topics.\r\n13) Result Publishing & Student Performance Module: This module publishes results to students. (i) display of marks and grades, (ii) access to feedback, (iii) performance trend visualization, and (iv) student dashboard result access.\r\n14) Exam & Question Paper Management Module: This module allows teachers to manage exams.(i) creation of MCQ and descriptive questions, (ii) question paper upload, (iii) definition of marks and grading rules, and (iv) mapping questions to learning outcomes.$$\n1) MCQ Detection & Auto-Checking Module\r\n2) Answer Sheet Upload & Management Module\r\n3) Rubric-Based Grading Module\r\n4) Descriptive Answer Evaluation Module\r\n5) Rechecking & Query Management Module\r\n6) Analytics & Performance Reporting Module\r\n7) Exam & Question Paper Management Module$$\n1) User & Role Management Module\r\n2) Image Preprocessing Module \r\n3) Optical Character Recognition (OCR) Module\r\n4) Cheating & Answer Similarity Detection Module\r\n5) Automated Feedback Generation Module\r\n6) Teacher Review (Human-in-the-Loop) Module\r\n7) Result Publishing & Student Performance Module$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-14 15:53:10","updated_at":"2026-03-23 19:16:05","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1071,"project_id":1466,"title":"Mentor AI","prob":"Many students face difficulty in finding the right tutor according to their subject, budget, language, and availability. Most students search tutors randomly on the internet or social media, which is time consuming and not reliable. There is no smart system that can analyze students\u2019 learning needs and recommend the most suitable tutors.\r\nThis project aims to solve this problem by creating an AI based platform where students can easily find, compare, and book qualified tutors from different parts of the world. The system will also analyze students\u2019 quiz results and recommend tutors according to their weak areas.","description":"MentorAI is an AI-based web platform that connects students with qualified tutors worldwide. Tutors will be able to create professional profiles where they can mention their subjects, experience, languages, availability, and hourly rates.\r\nStudents will be able to search tutors according to their learning needs such as subject, price, rating, and language. The platform will also include an AI Tutor Recommendation System that suggests the best tutors based on the student\u2019s preferences.\r\nIn addition, an AI Chatbot Assistant will help students by answering their questions and guiding them to find suitable tutors. The system will also analyze students\u2019 quiz results to identify weak areas and recommend tutors and learning paths for improvement.\r\nThe platform will include features such as session booking, live chat, video classes, notifications, trial sessions, assignment submission, and an admin dashboard to manage the system.$$\nUser Authentication Module\r\nHandles registration and login for students and tutors. Users create accounts using basic information like name, email, and password.\r\nStudent Module\r\nStudents can create and manage profiles, update interests, preferred language, and learning goals. They can search tutors, view profiles, ratings, reviews, book sessions, take quizzes, submit assignments, and communicate with tutors via chat or video.\r\nTutor Profile Module\r\nTutors create professional profiles with subjects, qualifications, experience, languages, hourly rate, and availability. They can manage schedules, accept or reject bookings, and view student feedback and ratings.\r\nTutor Search and Filter Module\r\nStudents can search for tutors by subject, language, price, rating, and availability, helping them quickly find suitable tutors.\r\nAI Tutor Recommendation Module\r\nUses AI to suggest tutors based on student preferences like subject, budget, language, and tutor ratings, providing a tailored list of recommended tutors.\r\nAI Chatbot Assistant Module\r\nActs as a virtual assistant, guiding students on using the platform, answering questions, helping find tutors, and providing quick support.\r\nQuiz Analysis Module\r\nStudents take quizzes, and the system analyzes results to identify weak areas. It recommends tutors and learning resources for improvement.\r\nBooking and Scheduling Module\r\nAllows students to book sessions. Tutors set availability, and students select suitable times. Automatic time zone adjustment ensures smooth scheduling across countries.\r\nCommunication and Video Class Module\r\nSupports chat, video classes, file sharing, and recorded sessions, enabling better interaction and learning.\r\nPayment Module\r\nManages secure payments through credit\/debit cards, PayPal, or other gateways. Stores payment records for tracking.\r\nAdmin Dashboard Module\r\nAdmins manage the platform, monitor users, sessions, and payments. They approve\/reject tutor registrations, block users violating rules, and generate reports to maintain quality and security.\r\nMulti-Language Support Module\r\nSupports multiple languages, enabling global users to use the platform easily.\r\nReviews and Rating Module\r\nStudents can rate tutors (1\u20135 stars) and write reviews after sessions. Ratings are visible on tutor profiles for transparency. Admins can monitor and remove inappropriate feedback.\r\nMessaging and Chat Module\r\nStudents and tutors communicate via messaging for questions about subjects, schedules, or sessions. Tutors can clarify lesson details before sessions start.$$\nIn MentorAI I shall develop the User Authentication Module, which covers user registration, login. will also implement the Tutor Profile Module and Tutor Search and Filter Module so students can view tutor details and find suitable tutors based on subject, rating, language and price. Additionally, I will develop the AI Chatbot Assistant Module, Messaging and Chatbot Conversation Module, Communication and Video Module and Multi-Language Support Module to improve accessibility for global users. I will also implement the Reviews and Rating Module, and contribute to the Admin Dashboard Module for managing tutors, reviews and monitoring user activities.$$\nIn MentorAI I shall develop the Student Module, which allows students to manage profiles, search tutors and book tutoring sessions. I will also implement the Booking and Scheduling Module with calendar system and automatic time zone adjustment for international users. Additionally, I will develop the AI Tutor Recommendation Module, Quiz Analysis Module, Payment Module with international payment methods. I will also contribute to the Admin Dashboard Module by implementing features related to student management, booking monitoring and payment records.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-13 21:34:04","updated_at":"2026-03-13 21:34:04","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1088,"project_id":1464,"title":"Personal Health Analytics & Motion Monitoring System for iOS Using Native Apple Frameworks","prob":"Many smartphone users want to understand their daily health and activity patterns, but existing applications are often complex or require additional wearable devices. Most applications focus on heart rate monitoring and wearable sensors instead of analyzing motion, balance, sleep environment, and activity patterns directly from the smartphone itself. This creates a gap for users who want simple and privacy-focused health insights directly from their iPhone.\r\nAnother issue is that many health applications provide advanced analytics, reports, and activity insights only in paid or premium versions, which limits access for many users.\r\nThis FYP solves the problem by developing a native iOS application that collects motion, activity, and environmental data from the iPhone using built-in sensors. The system analyzes the data using rule-based logic and presents daily, weekly, and monthly health insights through dashboards, charts, and alerts while keeping all user data stored locally on the device for privacy.","description":"Personal Health Analytics & Monitoring System is a native iOS application developed using Swift and Apple frameworks such as Core Motion. The application collects activity, motion, and environmental data from the iPhone sensors and stores the data securely on the device using local storage. The system analyzes the collected data using rule-based calculations to generate meaningful health insights. It provides daily summaries, weekly and monthly comparisons, health scores, and visual charts to help users understand their health patterns.\r\nAdditional modules such as gait and balance diagnostics, smart sleep environment monitoring, and health pattern detection provide deeper insights into the user\u2019s activity and behavior. The system also includes alerts, notifications, and report export features while maintaining full user privacy by keeping all data locally on the device.$$\nIn Personal Health Analytics & Monitoring System for iOS the modules are (i) Health Data Collection\r\n(ii) Activity & Motion Tracking\r\n(iii) Secure Local Data Storage\r\n(iv) Background Data Update\r\n(v) Daily Health Summary Dashboard\r\n(vi) Weekly & Monthly Comparison\r\n(vii) Rule-Based Health Alerts\r\n(viii) Health Score Calculation\r\n(ix) Charts & Graph Visualization\r\n(x) Alerts & Notifications\r\n(xi) Reports Export\r\n(xii) Privacy & User Control\r\n(xiii) Gait & Balance Diagnostic Module (Core Motion)\r\n(xiv) Smart Sleep Environment Monitor\r\n(xv) Health Pattern Finder$$\nI shall develop the Health Data Collection, Activity & Motion Tracking, and Gait & Balance Diagnostic modules. This includes using Core Motion framework to collect device motion data, analyzing walking patterns using accelerometer and gyroscope sensors, and implementing background data updates according to iOS system rules.$$\nI shall develop the user interface and analytics modules. This includes the daily dashboard, weekly and monthly comparisons, charts and graphs, health score calculations, smart sleep environment monitoring, health pattern detection, rule-based alerts, notifications, and exporting health reports in PDF format.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-17 02:44:08","updated_at":"2026-03-17 02:44:08","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1063,"project_id":1475,"title":"Smart Match: An Explainable, Transformer-Based Intelligent Job Portal for Context-Aware Resume\u2013Job Matching","prob":"Pakistan faces a graduate employability crisis, with over 16% educated unemployment in IT and engineering. Fresh graduates struggle to find relevant jobs because existing portals use simple keyword matching that misses semantic relationships. A candidate with Machine Learning experience is a perfect fit for an AI Developer role, but keyword systems fail to recognize this. Employers waste significant time screening unqualified applicants. Additionally, graduates have no structured guidance on which skills to develop for specific career paths, widening the skills gap further. SmartMatch solves all three problems: semantic job matching, explainable AI results, and a structured career roadmap with free upskilling resources.","description":"SmartMatch is a web-based intelligent recruitment platform. Job seekers register and either upload a PDF resume or build one using the built-in CV Builder, which generates a formatted PDF and feeds it directly into the AI pipeline. The Resume Parsing module extracts skills using SpaCy and NLTK. The Semantic Matching Engine uses Sentence-BERT with dynamic section-weighted embeddings and cosine similarity to compute match scores. For each match, SHAP explains which skills contributed positively or negatively. The Career Roadmap module lets students pick a target role, view a radar chart of skill gaps, and access free learning resources for each missing skill ranked by dynamic importance weights. Job data is sourced from employers, Rozee.pk and Mustakbil.com scrapers, and a 33,000-listing Kaggle dataset.$$\n1.CV Builder & User Interface Module: User registers\/selects role or fills CV Builder form. React.js interface with Bootstrap 5, JWT role-based authentication (Seeker\/Employer\/Admin). Multi-step CV Builder form generates a professionally formatted PDF via ReportLab, which auto-feeds into the AI matching pipeline \u2014 no manual upload needed. Job Seeker Dashboard displays colour-coded match cards (green >70%, amber 40\u201370%, red <40%), application tracking (Applied\/Saved\/Rejected), and skill-gap charts. Employer Interface allows job posting and AI-ranked candidate analytics.\r\n2.Resume Parsing & NLP Pipeline Module: User uploads a PDF resume. PDF text extraction via PyMuPDF. SpaCy and NLTK perform tokenisation, lemmatisation, and section segmentation (Skills, Experience, Education). Hybrid skill extraction combines SpaCy NER with taxonomy keyword matching, validated by SBERT semantic confirmation to reduce false positives. Structured skills list, experience level, education record stored in database and passed to Modules 3 and 4.\r\n3.Semantic Matching Engine & Explainability Module: Parsed resume sections and active job descriptions. SBERT (all-MiniLM-L6-v2) converts resume sections and job descriptions into 384-dimensional meaning-vectors. A dynamic section-weighted composite embedding is applied \u2014 Skills, Experience, and Education weights are automatically adjusted per job posting (not fixed) using two mechanisms: (1) SBERT role classification \u2014 the job description is matched against 15\u201320 IT role prototypes to assign a role-specific base weight profile (e.g., Data Science: Skills 0.50, Experience 0.35, Education 0.15; Entry-Level: Skills 0.50, Experience 0.20, Education 0.30); (2) keyword density adjustment \u2014 skill, experience, and education keyword counts in the job description apply a \u00b110% delta, normalised so weights always sum to 1.0. Cosine similarity between composite vectors produces the match score. SHAP then computes per-skill contribution scores, generating a natural-language explanation per match. Ranked jobs with scores and explanation e.g. \"Your 74% score is driven by Python (+0.15), but Docker (\u22120.08) is missing.\"\r\n4.Career Roadmap & Guided Upskilling Module: User selects a target IT career role. Role taxonomy of 15\u201320 IT careers (seeded from roadmap.sh) with skill requirements per role. SBERT-based gap scoring detects present, partial, and missing skill coverage. Missing skills are ranked dynamically \u2014 not by a fixed scheme \u2014 using a two-component importance score: (1) TF-IDF frequency across role-specific job corpus (Kaggle dataset) identifies skills distinctively important for the selected role; (2) SBERT semantic closeness to the role title captures conceptually central skills. Final priority = importance_score \u00d7 (1 \u2212 coverage_score), ensuring highly important absent skills rank first uniquely per user. Radar chart (Chart.js) with priority-ranked missing skills linked to free resources (YouTube, Coursera, edX). Progress tracked via Learning\/Learned status.\r\n5.Job Data Aggregation & Management Module: Job listings from Rozee.pk, Mustakbil.com, employers, and Kaggle LinkedIn dataset (33,000+, academic licence). Scrapy spiders crawl on a scheduled basis with deduplication filtering. Admin approval panel reviews scraped jobs before listing. SBERT embeddings are pre-computed for every ingested job at import time for real-time matching performance. Live Pakistan-relevant job database ready for semantic matching.$$\n1.CV Builder & User Interface Module: React.js interface, JWT auth, CV Builder form generating PDF via ReportLab, Job Seeker Dashboard with match cards and tracking, Employer Interface for posting and analytics.\r\n2.Resume Parsing & NLP Module: PDF extraction via PyMuPDF, SpaCy\/NLTK preprocessing, hybrid skill extraction using NER and taxonomy matching.\r\n3.Career Roadmap UI: Radar chart, skill gap display, importance-ranked free learning resource links and progress tracking interface.$$\n1.Semantic Matching & Explainability Module: SBERT embeddings, dynamic role-adaptive weights, SHAP per-skill reports, natural-language explanations, cold-start fallback.\r\n2.Career Roadmap & Skill Gap Engine: Role taxonomy (15\u201320 IT roles from roadmap.sh), TF-IDF + SBERT importance scoring, dynamically ranked missing skills, curated free resource library.\r\n3.Job Data Aggregation Module: Scrapy scrapers for Rozee.pk and Mustakbil.com, Kaggle dataset import, SBERT embedding pre-computation, admin approval panel.$$\n$$\n1.\tAN AI-POWERED JOB RECOMMENDATION SYSTEM FOR JOB\r\n2.\tJob Hunter$$\n1. Dynamic Section-Weighted Semantic Matching with SHAP: SBERT embeddings with role-adaptive weights automatically adjusted per job role using role classification and JD keyword analysis.$$\n2. CV Builder with Integrated AI Pipeline: Users build a formatted PDF resume on-platform, automatically submitted to the matching engine in one step.$$\n3. Career Roadmap with Importance-Ranked Upskilling: Missing skills ranked using TF-IDF corpus frequency combined with SBERT semantic importance per role, not a fixed scheme.","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-10 22:59:21","updated_at":"2026-03-10 23:04:29","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1067,"project_id":1478,"title":"AI-Based Non-Invasive Medical Analyzer (Real-Time Health Monitoring)","prob":"Many people lack access to affordable and continuous health monitoring systems. Traditional medical devices for measuring parameters like heart rate and oxygen saturation are often expensive and require hospital visits or trained personnel. This makes regular monitoring difficult, especially for patients with chronic conditions or those living in remote areas. The proposed project aims to address this problem by developing a low-cost, portable, non-invasive medical analyzer using an ESP32 microcontroller and optical PPG sensor. The system will monitor vital health parameters in real time and display them instantly, enabling individuals to easily track their health and detect potential issues early.","description":"This project aims to design and develop a smart non-invasive medical analyzer capable of monitoring vital health parameters such as heart rate and blood oxygen saturation (SpO\u2082) using an optical PPG sensor. The system is built around the ESP32 microcontroller, which reads the raw physiological signals from the sensor placed on a user\u2019s fingertip. The acquired signals are then processed to remove noise and extract useful information related to blood flow and pulse rate. A trained AI model or signal processing algorithm is used to analyze the processed signal and estimate the health parameters. The results are displayed on an OLED screen in real time, allowing the user to easily view their health status. The device is designed to be portable, low-cost, and easy to use, enabling continuous monitoring without the need for invasive medical equipment.$$\n\uf0b7 Sensor Module\r\n\uf0b7 Data Acquisition Module\r\n\uf0b7 Signal Processing Module\r\n\uf0b7 AI \/ Analysis Module\r\n\uf0b7 Display & User Interface Module\r\n\uf0b7 Power Management Module$$\nM Muneeb: will develop the Sensor and Data Acquisition Module. This module will handle the interfacing of the PPG sensor with the ESP32 microcontroller, reading physiological signals from the sensor through I2C communication, and collecting raw data for further processing.$$\nM Hammad : will develop the Display and Power Management Module. This module will handle the OLED display to show real-time health parameters such as heart rate and SpO\u2082. It will also manage the power supply system, including battery integration and voltage regulation, to ensure stable operation of the device.$$\nZain Riaz: will develop the Signal Processing and AI Analysis Module. This module will process the raw PPG signal received from the sensor, apply filtering techniques to remove noise, and implement machine learning or signal processing algorithms to estimate health parameters such as heart rate and oxygen saturation.$$\n$$\n$$\n$$\n","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-13 10:57:33","updated_at":"2026-03-13 11:30:41","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1095,"project_id":1480,"title":"AI-Based Real Estate Valuation Engine for Predicting Actual Property Market Prices Using Verified Local Data","prob":"The Pakistani real estate market suffers from severe \"Information Asymmetry.\" Publicly advertised \"Asking Prices\" on portals are often significantly inflated compared to actual \"Deal Prices\" settled privately, while official FBR\/DC rates often lag behind market reality. This opacity creates confusion for buyers, fosters mistrust between stakeholders, and increases the risk of financial fraud. Currently, no unified system exists to bridge the gap between these conflicting data sources. This project solves this problem by developing an AI-powered system that scientifically triangulates data from listings, government rates, and verified agent logs to estimate \"True Market Value,\" thereby democratizing market transparency and reducing financial risk for citizens.","description":"This project aims to develop an AI-powered real estate valuation system that estimates the realistic selling price of a property based on actual market behavior rather than relying solely on advertised asking prices. The system integrates three distinct data sources to improve valuation, accuracy and reliability.\r\nFirst, official government valuation data (FBR\/DC rates) will be used to establish a minimum price threshold, helping identify suspiciously low or potentially fraudulent listings. Second, publicly available asking price data (e.g., from Zameen.com) will define an upper price boundary, allowing the system to detect overpriced properties. Third, and most importantly, verified transaction data collected from real estate agents will be used to train the machine learning model, as it reflects actual deal prices and real market trends.\r\nBy combining these sources, the system will generate a realistic price range along with a confidence score, ensuring that predictions fall within meaningful market bounds. The project will initially focus on a selected housing society to ensure high-quality data and feasibility.\r\nThe system will include a complete end-to-end pipeline, including data collection, storage, preprocessing, machine learning-based valuation, model retraining, and a user-friendly web interface for querying property values.\r\nThe primary objective is to reduce information asymmetry in Pakistan\u2019s real estate market by providing transparent, data-driven property valuations, while maintaining feasibility within the scope of a one-year undergraduate final-year project.$$\nModule 1: Data Collection & Integration Engine\r\nThis module handles the collection and integration of property data from three sources:\r\n(1) official FBR\/DC rates (minimum threshold),\r\n(2) publicly available asking prices (e.g., Zameen.com) for upper bounds, and\r\n(3) verified transaction data from real estate agents.\r\nIt ensures all incoming data is standardized and stored in a unified format for further processing.\r\n\r\nModule 2: Data Wrangling & Verification\r\nThis module performs data cleaning, preprocessing, and validation. It removes inconsistencies, handles missing values, and applies verification logic such as:\r\n\u2022\tFlagging prices below FBR rates (potentially fake) \r\n\u2022\tFlagging prices above market asking trends (overpriced)\r\nThis ensures high-quality data for model training. \r\n\r\nModule 3: AI-Based Price Prediction Model\r\nThis is the core machine learning module. It trains regression models (e.g., Linear Regression, XGBoost) using verified real estate agent deal data to predict realistic property prices based on features like location, plot size, and amenities.\r\n\r\nModule 4: Price Range & Confidence Scoring\r\nThis module generates a price range instead of a single value by combining:\r\n\u2022\tML model prediction (center value) \r\n\u2022\tFBR rate (lower bound) \r\n\u2022\tAsking price trends (upper bound)\r\nIt also provides a confidence score based on data availability and consistency. \r\n\r\nModule 5: User Interaction \r\nThis module focuses on the frontend where users input property details and receive predicted price ranges along with confidence scores. It provides a simple and intuitive interface for querying property valuations.\r\n\r\n\r\n\r\nModule 6: Historical Price Trend & Analytics Module\r\nThis module provides users with historical price trends for their queried property type and location. It gives context to the predicted price by showing how prices have changed over time in that society\/block.$$\nI shall develop the modules User Interaction & Price Prediction and Historic Price Trend Analysis. My work will focus on designing user-friendly interfaces for property valuation queries and agent data submission. I will ensure smooth interaction between users and the system, allowing users to easily input property details and view predicted price ranges with confidence scores.$$\nI shall develop the modules Data Collection & Integration, Data Wrangling & Verification, and AI-Based Price Prediction Model. My work will focus on handling multi-source data, ensuring data quality, and developing machine learning models for accurate price prediction. I will also implement logic for generating price ranges using FBR and market asking data.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-29 14:57:29","updated_at":"2026-03-29 14:57:29","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1097,"project_id":1491,"title":"SabaqFlow: A Hybrid-Sync Adaptive Learning Ecosystem & SLO-Compliant Assessment Engine","prob":"Pakistan\u2019s education system relies heavily on rote memorization rather than conceptual learning, with traditional exams failing to accurately measure modern Student Learning Outcomes (SLOs). Furthermore, schools rarely evaluate students' cognitive aptitudes to guide them toward essential digital skills like AI or Web Development. The current digital learning experience is also passive and rigid; platforms like Canvas or Moodle operate on an \"always-online\" model, causing the learning process and student tracking to completely halt during internet outages.\r\n\r\nOn the educators' side, school teachers struggle to manually create balanced, SLO-compliant exams, while university instructors lack advanced tools like Retrieval-Augmented Generation (RAG) to instantly create quizzes from custom PDFs. These challenges highlight the urgent need for an AI-powered, hybrid-sync ecosystem that gives equal priority to seamless online streaming and robust offline access, solving these issues through automated RAG assessments, cognitive tracking, and interactive AI support.","description":"Introduction:\r\nSabaqFlow is a structured learning platform designed to bridge the digital divide. Unlike standard websites that just list content, SabaqFlow uses an \"Offline-Online Sync\" architecture. It helps teachers by providing an AI Exam Generator to create conceptual papers and a Student Tracker to monitor activities (e.g., checking if a student actually watched a lecture or skipped it). The system shifts focus from \"Ratta\" (memorization) to Conceptual Learning using AI to generate SLO based exams and in-video context aware quizzes. It also includes a central Admin Panel to manage the school's curriculum.\r\n\r\nObjectives:\r\nThe system aims to:\r\n\u2022\tEnable Hybrid Connectivity: Allow students to download videos and quizzes, solve them offline, and automatically update the data when internet is available.\r\n\u2022\tAutomate SLO-Compliant Assessments: SLO-compliant; single-click exam generation for SSC\/HSSC levels (enforcing a 30% recall \/ 70% conceptual split) and utilize advanced RAG pipelines for university-level conceptual quizzes directly from uploaded PDFs\/PPTs.\r\n\u2022\tImprove Student Interest: Add Gamification features like Streaks, Badges, and Certificates to keep students motivated.\r\n\u2022\tDetailed Student Tracking: Provide report to teachers with specific insights (e.g., \"fully watched vs. skipped video segments\").\r\n\u2022\tCentralized Administration: Empower admins to perform CRUD operations on curriculum content, users, and global settings.\r\n\u2022\tBridge the Language Gap: Provide a \"Bilingual Support\" for English-to-Urdu context.\r\n\u2022\tEnable Video-Based Quizzes: Automatically generate quizzes from video lectures that work during both Online Streaming and Offline Playback.\r\n\u2022\tEmpower Students with Explainable AI: A 24\/7 Explainable AI companion that answers queries with exact source attributions.$$\n1)\tTeacher Dashboard (Assessment & Tracking Engine):\r\nA web portal where teachers generate exams and monitor students.\r\no\tSLO-Exam Generator: Uses RAG to create conceptual exams mapped to Bloom\u2019s Taxonomy.\r\no\tDetailed Activity Logger: When a student synchronizes, the dashboard flags specific gaps (e.g., Alert: Ali started the Web Development roadmap but did not watch the 'CSS Grid' lecture assigned yesterday).\r\n\r\n2)\tSuper Admin Panel (Management):\r\n A centralized control center for school management.\r\no\tUser Management: Full CRUD capabilities for Teachers and Students.\r\no\tContent Governance: Ability to add\/update\/delete global courses (e.g., adding a new \"Generative AI\" module to the curriculum).\r\no\tSystem Analytics: High-level view of school-wide adoption and performance.\r\n\r\n3)\tStudent App (Mobile Application):\r\nA mobile application featuring Secure In-App Persistence.\r\no\tSmart Entry & Recommendation: Students can take a \"Cognitive Test \" where the system analyzes their aptitude and recommends specific Field Tracks (e.g., \u201cBased on your Logic Score, we recommend Data Science\u201d), Own Choice: Alternatively, students have the freedom to browse the full catalog and start any course of their own choice regardless of the recommendation.\r\no\tGamification & Certification: Streaks: Daily login and study tracking to build consistency, Smart Badges: Awards for milestones (e.g., \"Quiz Master,\" \"Night Owl\"), Auto-Certification: Generates a verifiable PDF certificate upon completing a \"Trending Roadmap\" (e.g., Python Basics).\r\no\tUniversal Quiz Engine (Online\/Offline): Online Mode: As the student streams a video, the system reads the transcript to present real-time concept checks, Offline Mode: When downloading a video, the system saves the transcript, generates a quiz via the backend and saves it to local storage. This allows the student to have the exact same interactive quiz (based on real time content) experience without internet using local storage. Whenever student connects with the internet; progress updates on cloud storage.\r\no\tDigital Exam Hall: Allows students to solve Teacher-assigned Model Papers digitally.\r\n\r\n4)\tAI Companion (The Smart Assistant):\r\nAn integrated AI chatbot (Online) acting as a 24\/7 tutor where students can ask conceptual queries or get help with problems. Key Feature: Unlike standard bots, it provides Source Attribution with every answer (e.g., \"Reference: Class 9 Computer Science Textbook, Chapter 3\" or \"Related Video: Intro to Arrays at 04:20\"), ensuring students can verify the information.\r\n\r\n5)\tThe Language Bridge (Bilingual Support):\r\nA cross-module layer that ensures roadmaps and explanations are available in both English and Urdu.$$\nAhmer Talal shall develop the Backend Intelligence, Teacher & Admin Ecosystem. My deliverables:\r\no\tSuper Admin & Teacher Portal: Developing the complete React.js Frontend (dashboards, charts, exam generation) for Admin management and Teacher.\r\no\tThe Exam Engine: Python pipeline to index textbooks and generate SLO-compliant exams.\r\no\tVideo & Transcript Intelligence: Backend logic to handle YouTube API requests, fetch metadata, and extract transcripts for the AI engine.\r\no\tAI Assistant Integration: Integrating the LLM API for student queries.\r\no\tCognitive Analysis Logic: The backend algorithm that processes the \"Cognitive Audit\" inputs and maps them to recommended curriculum tracks\r\no\tDatabase Architecture: Managing the Firebase Structure for multi-role access (Admin\/Teacher\/Student).$$\nMuhammad Obaidullah shall develop the Student Mobile Experience & Offline Architecture. My deliverables:\r\no\tCore Mobile Interface (Flutter): Developing the complete mobile frontend structure and responsive UI.\r\no\tVideo Player & Download Interface: Implementing the video player UI and logic for secure in-app downloads.\r\no\tBilingual Support UI: Urdu\/English toggle logic.\r\no\tGamification & Digital Exam UI: Implementing logic for Badges, Streaks, Certificate Generation, and the Digital Paper interface.\r\no\tAptitude Test Interface: The interactive screen for the Cognitive Audit and displaying the recommended roadmap.\r\no\tSecure Local Storage: SQLite logic to store videos and track offline activity (which lecture was watched vs. skipped) for the sync payload.$$\n$$\nSMART SCHOOL (School Management System)$$\n1) Detailed Offline Tracking: Unlike standard apps, our system tracks specific offline actions (watched vs. skipped) and reports them to the teacher upon sync.$$\n2) SLO-Driven Assessment Matrix: Enforces a 30% Recall \/ 70% Concept split in automated exams.$$\n3) Universal Context-Aware Logic: Extracts video transcripts to provide interactive quizzes whether the student is streaming online or watching downloaded content offline.","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-31 20:33:58","updated_at":"2026-03-31 20:33:58","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1083,"project_id":1499,"title":"VIZYON: AI-Based University Admission Recommendation and Guidance System for Turkey","prob":"Students who wish to pursue higher education in Turkey often face difficulty in finding accurate and organized information about universities, admission requirements, eligibility criteria, deadlines, scholarships, and application preparation. The information is scattered across multiple sources, making the process time-consuming and confusing. Many students rely on expensive consultants or repeatedly search for the same details without proper planning. Additionally, students struggle to understand which universities best match their academic performance and how to prepare application documents such as essays and Statements of Purpose. This Final Year Project aims to solve these problems by providing an AI-based advisory system that analyzes student academic progress, suggests suitable universities, assists in essay preparation, and organizes the admission process through a simple planner, making admission guidance more accessible and structured.","description":"VIZYON is an AI-based university admission guidance system designed to assist students who plan to apply to universities in Turkey. The system focuses on helping students find suitable universities, plan their admission journey, and prepare application-related content in an organized and structured manner. Instead of automating applications or claiming real-time integrations, VIZYON works as a decision-support and advisory platform.\r\n\r\nThe system allows students to input their academic performance, including grades, marks, and educational background. Based on this data, VIZYON analyzes the student\u2019s academic progress and suggests universities and programs that align with their profile using rule-based logic and basic machine learning techniques. Students can explore universities through a college-finding feature, apply multiple filters, and save their shortlisted options in a built-in planner so they do not need to search repeatedly.\r\n\r\nVIZYON uses a structured dataset that is built using information available on publicly accessible academic websites. Basic details such as university names, locations, and general institutional information are collected from these public sources and organized within the system\u2019s internal database. This dataset acts as the main reference that VIZYON uses when suggesting universities and programs to students based on their academic profile. To ensure that the information remains useful and reasonably up to date, the system follows a hybrid data management approach. This enables the dataset to be periodically reviewed and updated whenever changes are detected at website, instead of relying on static data.\r\n\r\nAdditionally, VIZYON provides advisory-level scholarship awareness by displaying general opportunities and eligibility hints. The system also includes a planner that organizes tasks, deadlines, and preparation steps, helping students track progress and identify pending actions. Overall, VIZYON simplifies the admission preparation process while keeping the system realistic, ethical, and feasible as a Final Year Project.$$\n1. Student Academic Analysis Module\r\nThis module collects and analyzes student academic data such as grades, marks, degree background, and test scores. It evaluates the overall academic progress of the student and prepares the data for university recommendation and eligibility analysis.\r\n2. University Finding & Recommendation Module\r\nThis module enables students to search and discover universities in Turkey through the data from the Database which has been collected from the official websites having public data. It uses rule-based logic and machine learning techniques to suggest universities and programs that match the student\u2019s academic profile, preferences, and eligibility conditions.\r\n3. Eligibility Evaluation Module\r\nThis module checks basic eligibility criteria such as minimum grades, degree relevance, and general test requirements. It categorizes universities as suitable, partially suitable, or not suitable, helping students make informed decisions.\r\n4. Scholarship Awareness Module\r\nThis module displays general scholarship opportunities and provides advisory-level eligibility hints. It does not guarantee funding or integrate with official scholarship portals.\r\n5. Admission Planner Module\r\nThis module allows students to save searched universities, plan tasks, track deadlines, and organize all admission-related activities in one place so repeated searching is avoided.\r\n6. Progress Tracking Module\r\nThis module tracks student preparation progress, highlights completed and pending tasks, and provides non-binding feedback on readiness level.$$\nIn VIZYON, I shall develop the Student Academic Analysis Module, which collects and analyzes student academic data such as grades, marks, degree background, and test scores to evaluate the student\u2019s academic profile.I will also develop the University Finding & Recommendation Module, which enables students to search and discover universities in Turkey using data stored in the system database compiled from publicly accessible sources. This module will use rule-based logic and basic machine learning techniques to suggest universities and programs that match the student\u2019s profile and preferences.Additionally, I will work on the Eligibility Evaluation Module, which checks basic admission criteria such as minimum grades and degree relevance, and categorizes universities as suitable, partially suitable, or not suitable.$$\nIn VIZYON, I shall develop the Scholarship Awareness Module, which displays general scholarship opportunities and provides advisory-level eligibility hints to help students explore possible funding options. This module will present informational guidance rather than integrating directly with official scholarship systems.I will also implement the Admission Planner Module, which allows students to save searched universities, plan admission-related tasks, track deadlines, and organize their application preparation activities in one place so repeated searching is avoided.In addition, I will develop the Progress Tracking Module, which monitors the student\u2019s preparation progress by highlighting completed and pending tasks and providing general feedback on readiness level.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-15 20:50:23","updated_at":"2026-03-15 20:50:23","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1057,"project_id":1508,"title":"Style With Us \u2013 An AI-Based Personal Fashion Styling and Recommendation Mobile Application","prob":"Online fashion shoppers often struggle to determine whether a particular clothing style or color will suit their body structure and skin tone before making a purchase. This uncertainty leads to dissatisfaction, poor buying decisions, and high product return rates. Most existing platforms only display products and lack personalized styling assistance. There is no intelligent mobile solution that guides users according to their physical attributes. This FYP solves this issue by providing AI-based body type detection and skin tone analysis to recommend suitable outfit styles and color palettes tailored to each user","description":"The user registers and uploads an image through the mobile application. The system processes the image to detect body type and skin tone using AI-based image analysis techniques. Based on these attributes, personalized outfit styles and color recommendations are generated. Users can browse suggested clothing options and evaluate recommended combinations before proceeding with any purchase decision. The application serves as an intelligent styling assistant to improve decision-making and reduce incorrect purchases.$$\n. User Management Module Handles registration, login\/logout, profile management, and secure image upload. This module ensures that user data is stored securely and that images uploaded for analysis are associated with the correct user profile.\r\n2. Body Type Analysis Module Performs image preprocessing and classifies the user's body structure for styling recommendations. It utilizes computer vision techniques to analyze the uploaded image and categorize the user's physique (e.g., hourglass, rectangular) to determine the most flattering clothing cuts.\r\n3. Skin Tone Detection Module Detects skin tone and generates appropriate color palette suggestions. By analyzing facial regions in the uploaded image, this module identifies the user's undertone (warm, cool, or neutral) and recommends specific colors that enhance their appearance.\r\n4. Outfit Recommendation Module Combines analysis results to recommend clothing styles and outfit combinations. This is the core logic engine that takes the outputs from the Body Type and Skin Tone modules to filter and suggest specific apparel that matches both criteria.\r\n5. Product Browsing Module Displays recommended clothing items and allows exploration of outfit categories. It serves as the front-end catalog where users can view the curated suggestions, see details about the items, and visualize how different pieces work together.$$\n1. User Management Module Handles registration, login\/logout, profile management, and secure image upload. This module ensures that user data is stored securely and that images uploaded for analysis are associated with the correct user profile.\r\n2. Body Type Analysis Module Performs image preprocessing and classifies the user's body structure for styling recommendations. It utilizes computer vision techniques to analyze the uploaded image and categorize the user's physique (e.g., hourglass, rectangular) to determine the most flattering clothing cuts.\r\n3. Skin Tone Detection Module Detects skin tone and generates appropriate color palette suggestions. By analyzing facial regions in the uploaded image, this module identifies the user's undertone (warm, cool, or neutral) and recommends specific colors that enhance their appearance.\r\n4. Outfit Recommendation Module Combines analysis results to recommend clothing styles and outfit combinations. This is the core logic engine that takes the outputs from the Body$$\n\uf0b7 Image Preprocessing Pipeline:\r\n\uf0b7 Developing the initial processing stage using Python and OpenCV.\r\n\uf0b7 Implementing algorithms for noise reduction (Gaussian blur), resizing, and Background Removal to isolate the user's body from the surroundings for accurate detection.\r\n\uf0b7 Body Type Detection Logic:\r\n\uf0b7 Feature Extraction: coding logic to detect key body landmarks (shoulders, waist, and hips) using contour detection or pose estimation libraries.\r\n\uf0b7 Ratio Calculation: Implementing mathematical formulas to calculate ratios between body parts (e.g., shoulder-to-waist ratio).\r\n\uf0b7 Classification: Developing the decision tree that maps these ratios to standard body shapes (Hourglass, Pear, Rectangle, Inverted Triangle, Apple).$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-06 00:58:21","updated_at":"2026-03-06 00:58:21","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1073,"project_id":1502,"title":"Smart Hydro Power Monitoring and Forecasting System (SHPMFS)","prob":"Hydroelectric operations under PEDO play an important role in regional energy supply but current monitoring practices rely on fragmented data sources and manual spreadsheet records which limit centralized monitoring and real time operational analysis across powerhouses. Electricity generation and hydrology data are often stored separately which makes it hard to analyze generation efficiency and long term operational trends. Forecasting is mostly reactive and not based on systematic analysis of historical data and operational anomalies are often detected late due to the absence of automated monitoring systems which reduces operational efficiency and delays decision making. The proposed Smart Hydro Power Monitoring and Forecasting System aims to address these challenges by providing centralized data management along with generation analytics dashboards and AI based electricity forecasting and anomaly detection with analytical root cause insights and automated reporting to support more efficient and data driven hydroelectric operations.","description":"The Smart Hydro Power Monitoring and Forecasting System is a data driven platform designed to improve hydroelectric monitoring through centralized data integration and predictive analytics. The system collects electricity generation and hydrology data from hydro powerhouses and stores it in a structured database which is then analyzed through interactive dashboards and correlation analysis along with forecasting models and automated reporting tools. An AI Forecasting Engine uses machine learning models to predict short term electricity generation based on historical generation data and hydrological patterns. The system also includes an anomaly detection module that continuously monitors operational data to identify unusual patterns and provides analytical root cause insights to assist users in identifying potential operational issues. Administrators can set generation targets for each powerhouse and the system automatically compares actual values with those targets to evaluate performance. The main objective is to transform hydroelectric monitoring from manual and reactive processes into a centralized and predictive system that improves operational visibility and forecasting capability along with decision making efficiency.$$\n1. Powerhouse Data Collection and Storage System\r\nA centralized database system that collects electricity generation and water inflow and outflow data from hydro powerhouses and supports historical data storage and integration of legacy datasets such as Excel or CSV files.\r\n\r\n2. Generation Analytics Dashboard\r\nAn interactive dashboard that visualizes electricity generation trends across powerhouses using charts and performance indicators for daily and weekly and monthly analysis.\r\n\r\n3. Hydrology Generation Correlation Engine\r\nAnalyzes the relationship between water inflow and electricity generation to compute efficiency metrics and identify utilization patterns.\r\n\r\n4. AI Forecasting Engine\r\nUses suitable machine learning models to forecast electricity generation for short term operational periods using historical generation and hydrology data.\r\n\r\n5. Anomaly Detection and Root Cause Analysis\r\nContinuously monitors operational data to detect abnormal generation or water flow patterns and provides estimated root cause insights based on data patterns to assist users in identifying potential operational issues.\r\n\r\n6. Target Setting and Performance Monitoring\r\nAllows administrators to define generation targets for each powerhouse and automatically compares actual generation values with targets to evaluate performance.\r\n\r\n7. Generation Forecast Dashboard\r\nDisplays predicted electricity generation values along with historical trends to help users understand expected generation patterns.\r\n\r\n8. Automated Weekly Reporting System\r\nGenerates automated weekly reports summarizing generation trends and forecasting results along with hydrology insights and detected anomalies.\r\n\r\n9. User Access Control and Security\r\nImplements role based authentication and maintains activity logs to ensure secure and accountable access to all operational data.$$\nIn SHPMFS I shall develop the following modules.\r\n(i) Powerhouse Data Collection and Storage System which includes centralized data collection and historical data storage and legacy dataset integration.\r\n(ii) Generation Analytics Dashboard which includes trend visualization and performance indicators for daily and weekly and monthly analysis.\r\n(iii) Hydrology Generation Correlation Engine which includes efficiency metrics computation and utilization pattern identification.\r\n(iv) Target Setting and Performance Monitoring which includes generation target definition and automatic performance comparison.\r\n(v) User Access Control and Security which includes role based authentication and activity log maintenance.$$\nIn SHPMFS I shall develop the following modules.\r\n(i) AI Forecasting Engine which includes machine learning based electricity generation forecasting using historical and hydrology data.\r\n(ii) Anomaly Detection and Root Cause Analysis which includes abnormal pattern detection and estimated root cause insights.\r\n(iii) Generation Forecast Dashboard which includes predicted generation visualization and historical trend display.\r\n(iv) Automated Weekly Reporting System which includes automated report generation summarizing generation trends and forecasting results and hydrology insights and detected anomalies.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-14 05:44:32","updated_at":"2026-03-14 05:44:32","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}
No. Project Title Reg. No. Name Date & Time Evaluation Status
1. SmartStudyInstructor: AI-Powered Hybrid Courses Platform with Adaptive, Personalized, Interactive Lecture Videos, Attendance, Engagement, and Assessments CIIT/SP23-BCS-011/WAH
CIIT/SP23-BCS-093/WAH
 RAJA HARIS KHUDADAD
MUHAMMAD ADAN KHURSHID
 Done Accepted  
2. DevSynx: AI-ERP for software houses with team management, client portal, real-time tracking, and automated workflow. CIIT/SP23-BSE-029/WAH
CIIT/SP23-BSE-056/WAH
 MARYAM JUNAID
SYED HUSNAIN RAZA GILLANI
 Feb 20, 2026
09:10 AM 		Revision Pending 
3. AI-Based Intelligent Resume Screening and Job Matching System CIIT/SP23-BSE-028/WAH
CIIT/SP23-BSE-043/WAH
 ADIL QAYYUM
JAWAD RAFIQUE
 Feb 20, 2026
09:30 AM 		Revision Pending 
4. Flocksy:A Unified Social Media Platform Offering Social Networking for Adult and Safe Learning for Children. CIIT/SP23-BCS-082/WAH
CIIT/SP23-BCS-090/WAH
 MUTTI UR REHMAN
MUHAMMAD SUFYAN KHAN
 Feb 20, 2026
09:50 AM 		Revision Pending 
5. AI-Based Recruitment Platform for Resume Analysis, Real-Time AI Interviews, and Candidate Evaluation. CIIT/SP23-BCS-031/WAH
CIIT/SP23-BCS-035/WAH
 SAMINA KHAN
KASHAF ZAHRA
 Feb 20, 2026
10:10 AM 		Revision Pending 
6. FINSIGHT Financial Intelligence System for Investment & Growth Tracking CIIT/SP23-BSE-014/WAH
CIIT/SP23-BSE-057/WAH
 AHMED MALIK
MUHAMMAD HASSAAN SHAHID
 Feb 20, 2026
10:30 AM 		Pending  
7. Smart AI-Assisted Team Task Management Platform for Software Houses with Progress Tracking and Reporting CIIT/SP23-BSE-005/WAH
CIIT/SP23-BSE-036/WAH
 HAIDER ALI MOION
MUHAMMAD MUZAMMIL TUFAIL
 Feb 20, 2026
10:45 AM 		Revision Pending 
8. Tez rider CIIT/SP23-BSE-033/WAH
CIIT/SP23-BSE-042/WAH
 AYESHA NADEEM
AROOBA AHSAN
 Feb 20, 2026
11:05 AM 		Revision Pending 
9. Doctor AI: An Intelligent AI-Driven Web Platform for Mental Health Support and Emotional Analysis CIIT/SP23-BCS-077/WAH
CIIT/SP23-BCS-105/WAH
 ZINNIA MELISSA
WALEED AHSAN KHAN
 Feb 20, 2026
11:25 AM 		Pending  
10. AutoGrade CIIT/SP23-BCS-047/WAH
CIIT/SP23-BCS-049/WAH
 MUHAMMAD USMAN
MUHAMMAD ARYAN AFZAL
 Feb 20, 2026
11:45 AM 		Pending  
11. Mentor AI CIIT/SP23-BCS-109/WAH
CIIT/SP23-BCS-113/WAH
 HANIYA FAYYAZ
NADIA ALI
 Feb 20, 2026
12:15 PM 		Revision Pending 
12. Personal Health Analytics & Motion Monitoring System for iOS Using Native Apple Frameworks CIIT/SP23-BCS-065/WAH
CIIT/SP23-BCS-069/WAH
 ALI IBRAHIM
USAMA ZAFAR
 Feb 23, 2026
08:30 AM 		Revision Pending 
13. A Template-Based E-Commerce Website Builder with Integrated Payment System CIIT/SP23-BCS-001/WAH
CIIT/SP23-BCS-026/WAH
 HISSAM MOHUDIN
SUFIAN AMIN
 Feb 23, 2026
08:50 AM 		Pending  
14. ARCHON: An Autonomous, Privacy-First AI Data Scientist with Evolutionary Model Selection and Logic-Based Reasoning. CIIT/SP23-BCS-076/WAH
CIIT/SP23-BCS-080/WAH
 RANA MUHAMMAD TAYYAB
MUHAMMAD SHAHID
 Feb 23, 2026
09:10 AM 		Revision Pending 
15. BOATMART – The All-in-One Solution for E-Commerce Module B CIIT/SP23-BCS-072/WAH
CIIT/SP23-BCS-103/WAH
 MUHAMMAD TALHA YASIN
SYEDA WAHEEDA GILLANI
 Feb 24, 2026
08:30 AM 		Pending  
16. Smart Match: An Explainable, Transformer-Based Intelligent Job Portal for Context-Aware Resume–Job Matching CIIT/SP23-BCS-039/WAH
CIIT/SP23-BCS-057/WAH
 MUHAMMAD EHSAN MUMTAZ
HAFFAZ UR REHMAN
 Feb 24, 2026
08:50 AM 		Pending  
17. AI-Based Non-Invasive Medical Analyzer (Real-Time Health Monitoring) CIIT/FA22-BCS-100/WAH
CIIT/FA22-BCS-035/WAH
CIIT/FA22-BCS-080/WAH
 Zain Riaz
Muhammad Muneeb
Muhammad Hammad
 Feb 24, 2026
09:10 AM 		Pending  
18. AI-Based Real Estate Valuation Engine for Predicting Actual Property Market Prices Using Verified Local Data CIIT/SP23-BCS-025/WAH
CIIT/SP23-BCS-092/WAH
 MASOOD-UR-REHMAN
MUHAMMAD AHMAD
 Feb 24, 2026
09:30 AM 		Revision Pending 
19. AI-Powered Smart Classroom with Virtual Student Avatars for Real-Time Question Generation CIIT/SP23-BCS-063/WAH
CIIT/SP23-BCS-101/WAH
 MUHAMMAD TALHA
NAUMAN AHMED
 Feb 24, 2026
09:50 AM 		Revision Pending 
20. SabaqFlow: A Hybrid-Sync Adaptive Learning Ecosystem & SLO-Compliant Assessment Engine CIIT/SP23-BCS-041/WAH
CIIT/SP23-BCS-112/WAH
 AHMER TALAL
MUHAMMAD OBAIDULLAH
 Feb 24, 2026
10:10 AM 		Revision Pending 
21. SoloVision: Assistive AI Glasses for Visually Impaired and Color Blind CIIT/SP23-BCS-027/WAH
CIIT/SP23-BCS-036/WAH
 SAJJAD
MUHAMMAD FAROOQ NAEEM
 Feb 24, 2026
10:30 AM 		Revision Pending 
22. VIZYON: AI-Based University Admission Recommendation and Guidance System for Turkey CIIT/SP23-BSE-019/WAH
CIIT/SP23-BSE-020/WAH
 MARYAM SHAHID
JAVERIA NASIM
 Feb 24, 2026
10:45 AM 		Revision Pending 
23. Automatic Solar Tracking And Cleaning System With Real-Time Energy Monitoring CIIT/SP23-BCS-099/WAH
CIIT/SP23-BCS-108/WAH
 UMER BASHIR
ALEENA AYYAZ
 Feb 27, 2026
10:15 AM 		Revision Pending 
24. Smart Cosmetic E-Commerce System with Machine Learning Based Fraud Detection CIIT/SP23-BCS-070/WAH
CIIT/FA22-BCS-021/WAH
 HIFSA EMAN
Eman Abid
 Feb 27, 2026
10:35 AM 		Revision Pending 
25. Style With Us – An AI-Based Personal Fashion Styling and Recommendation Mobile Application CIIT/FA22-BSE-112/WAH
CIIT/FA22-BSE-082/WAH
 Saif Ur Rehman
Ubaid Ahmed Khan
 Feb 27, 2026
10:55 AM 		Revision Pending 
26. Smart Hydro Power Monitoring and Forecasting System (SHPMFS) CIIT/SP23-BSE-058/WAH
CIIT/SP23-BSE-059/WAH
 FATIMA LIAQAT
MIRAYE AZIZ
 Feb 27, 2026
11:15 AM 		Revision Pending 
Committee: 3
Team Members:
Dr. Jamal Hussain Shah (HEAD)
Dr. Mamuna Fatima
Ikram Ul Haq
Ms. Sania Umer
Venue: CS Conference Room
Remarks: Please ensure that you arrive on time for the Proposal Presentation. Your punctuality is highly appreciated, as it reflects professionalism and helps us begin the session as scheduled. Kindly bring your laptop along with the necessary connector or adapter if your device does not support an HDMI cable. Thank you.

1{"id":1053,"project_id":1431,"title":"Hate Speech & Violence Detection in Broadcast Media","prob":"Traditional media monitoring systems fail to detect Implicit Hate Speech, specifically Irony and Sarcasm. For example, when a speaker mockingly praises a controversial figure (e.g., \"He is a genius for ruining us\"), standard filters see \"genius\" and classify it as positive. This \"Context Blindness\" allows hate speech to bypass censorship. Furthermore, existing systems are often limited to monolingual text. This project solves this by building an advanced, multilingual (Urdu & English) AI model that processes multimodal inputs (video, voice, text) to understand the incongruity between words and reality, ensuring accurate detection of hidden hate speech in real-world media.","description":"The system is an advanced, multimodal NLP pipeline designed to detect subtle forms of hate speech in broadcast media. Following an industrial standard, the workflow is: \r\nInput -> Pre-processing -> Understanding -> Classification. \r\nFirst, the system ingests multimodal data (video, voice, text). The Pre-processing stage extracts audio, identifies active speakers via Speaker Diarization, and transcribes the speech into text. \r\nSecond, the Understanding stage uses a Context-Extraction Engine to identify the \"Target\" (entity) and retrieves its reputation, while the Irony Detection Module analyzes \"Semantic Incongruity\". Finally, a Multilingual Transformer Model (mBERT\/XLM-RoBERTa) classifies the Urdu\/English content as Safe, Hate Speech, or Sarcastic\/Ironic. The results, along with proper AI explanations (XAI), are visualized on an interactive dashboard with graphs.$$\n1. Multimodal Data Ingestion & Pre-processing Module: \r\n\r\nThis module prepares raw broadcast feeds for analysis. It uses FFmpeg for audio extraction and PyAnnote for Speaker Recognition (Diarization) to track who is speaking. OpenAI Whisper then transcribes the voice into text. Finally, it performs text normalization, noise removal, and handles specific broadcast artifacts. \r\n\r\n2. Entity & Context Knowledge Module: \r\n\r\nThis module answers \"Who is being talked about?\". It uses Named Entity Recognition (NER) to extract the names of politicians or groups. It references a \"Context Knowledge Base\" to assign a reputation score (e.g., \"Controversial\") to the entity, which is essential for determining whether praise is genuine or ironic. \r\n\r\n3. Semantic Incongruity & Irony Detection Module: \r\n\r\nThis is the novel research component. It performs Aspect-Based Sentiment Analysis to check the sentiment towards the specific entity. It then calculates the contrast between the Sentiment (e.g., \"Great\") and the Context (e.g., \"Criminal\"). If a high contrast is found, it flags the statement as Irony\/Sarcasm. \r\n\r\n4. Multilingual Classification & Explainable AI (XAI) Module: \r\n\r\nThis module uses a Multilingual Transformer-based model (mBERT or XLM-RoBERTa) fine-tuned on hate speech and sarcasm datasets to natively support both Urdu and English. It incorporates Explainable AI (LIME\/SHAP) to provide proper reasoning for its decisions. To handle the computational load, training is executed on Cloud-based GPUs (Google Colab\/Kaggle T4 GPUs). The module predicts the final class: Safe, Hate Speech, Irony, or Violence. \r\n\r\n5. Command Center & Visualization Dashboard: \r\n\r\nA highly attractive, user-friendly web interface that logs all analyzed statements. It displays structured outputs using graphical charts (e.g., Chart.js\/D3.js) and timeline visualizations of speaker violations. It highlights detected hate speech in Red and Irony\/Sarcasm in Yellow, alongside the XAI explanations showing exactly why the AI flagged the content.$$\nI will be responsible for the Context-Aware Logic and Classification Modules. My work focuses on the backend AI intelligence, where I will develop algorithms for Irony Detection to identify semantic incongruity. I will handle the fine-tuning of the Multilingual BERT (mBERT) Model for Urdu and English classification. Additionally, I will implement the Explainable AI (XAI) overlays to ensure the model provides proper explanations for its classification results.$$\nI will be responsible for Application Interface and Data Management. I will develop the Multimodal Pre-processing Module, integrating tools for audio extraction, speaker recognition (diarization), and transcription to clean the raw input. Furthermore, I will build an attractive Web-based Reporting Dashboard featuring structured data visualizations (graphs and charts) and manage the backend database to store logs and evidence.$$\n$$\n$$\n$$\n$$\n","comments":"","isDraft":1,"status":1,"created_at":"2026-03-02 08:49:22","updated_at":"2026-03-25 23:19:07","isReviewDraft":1,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1054,"project_id":1476,"title":"SmartStudyInstructor: AI-Powered Hybrid Courses Platform with Adaptive, Personalized, Interactive Lecture Videos, Attendance, Engagement, and Assessments","prob":"Universities increasingly offer hybrid learning, yet existing platforms lack an integrated mechanism to convert teacher materials into structured lecture delivery with measurable participation and learning analytics. Teachers cannot consistently scale lecture delivery across multiple sections while tracking student engagement and learning progress. Students learning remotely often receive passive content without course-specific support, and institutions lack a unified view of attendance, engagement, and assessment performance.\r\n\r\nThis project proposes a hybrid course platform where teachers upload course documents and learning objectives, and the system builds a course-specific knowledge base using Retrieval-Augmented Generation (RAG). It generates lecture scripts and produces lecture videos with slides and a talking avatar (lip-sync), delivered via a student application. During playback, the system records attendance and engagement signals and conducts MCQ-based quizzes\/exams with automated grading. A student learning model updates mastery and confidence to support explainable recommendations and targeted practice.","description":"SmartStudyInstructor is an AI-powered hybrid course delivery platform (web + mobile) that converts teacher-uploaded material into structured lecture delivery with integrated monitoring and assessment. Teachers create a course, define outcomes\/learning objectives, and upload PDFs\/PPTs. The system performs document processing and builds a searchable course knowledge base using RAG (text extraction, chunking, embeddings, vector retrieval). Using retrieved course content, it generates a lecture script and narration. The system then produces a lecture video in which slides appear as the background and an avatar appears on the side with voice narration and lip synchronization, enabling students to consume lectures remotely through the mobile application.\r\n\r\nWhile the lecture video plays, the student application records participation through watch-time and a lightweight, privacy-aware camera-based engagement signal (face presence\/looking-at-screen) sampled periodically during the session. This provides measurable lecture attendance and engagement indicators without storing raw video. Students can ask questions during or after the lecture via a course-specific Q\/A feature; the system answers using retrieved material only and can show citations\/snippets to reduce hallucinations.\r\n\r\nAfter each lecture, the platform conducts MCQ-based quizzes\/exams, auto-grades results, and updates a student learning model (topic mastery, confidence, pace, and engagement trend). Teachers monitor attendance, engagement, performance, and explainable recommendations through a dashboard, allowing them to manage multiple sections with reduced repetitive workload. Optional extensions (OCR checking, plagiarism) are documented as future enhancements, while evaluation focuses on a working, demonstrable hybrid pipeline.$$\nModule 1: Authentication & Role Management\r\nProvides secure login\/logout and role-based access for Teacher and Student (and optional Admin). Controls access to course content, lecture generation, monitoring, and assessments.\r\n\r\nModule 2: Knowledge Intelligence (RAG-Based) Module\r\nPerforms document ingestion (PDF\/PPT), text cleaning, chunking, embeddings generation, and vector database indexing. Supports retrieval of top-k relevant content for lecture generation and Q\/A, with optional citations\/snippet references.\r\n\r\nModule 3: Teacher Portal & Course Management\r\nAllows teachers to create courses\/sections, enroll students, define learning objectives (OBE), upload material, generate\/publish lectures and quizzes, and monitor class analytics.\r\n\r\nModule 4: Lecture Script + Slide Generation Module\r\nGenerates an objective-aligned lecture script and narration script. Produces slides automatically (titles\/bullets) and timestamps for synchronization with narration.\r\n\r\nModule 5: Video Generation Module (Slides + Avatar + Lip-Sync)\r\nGenerates narration audio using free TTS and produces the lecture MP4 by combining slide background with a side avatar panel. Lip-sync is applied to the avatar using a lightweight pipeline; a fallback mode (static avatar + subtitles) ensures video generation reliability.\r\n\r\nModule 6: Student App (Hybrid Lecture Delivery + Q\/A)\r\nDisplays scheduled\/past lectures, streams videos, supports resume playback, and allows course-specific Q\/A during lecture.\r\n\r\nModule 7: Attendance & Engagement Monitoring Module\r\nComputes attendance using watch-time thresholds and periodically sampled camera-based engagement signals. Stores only numeric metrics (no raw video) and supports teacher override.\r\n\r\nModule 8: Assessment & Student Modeling + Analytics\/Explainability\r\nGenerates MCQ quizzes\/exams aligned to objectives, auto-grades results, updates student mastery\/confidence, and presents dashboards for teacher\/student. Maintains explainability logs for recommendations (e.g., weak topic flagged, tier selection).$$\nI will develop the student-facing application and user experience across hybrid learning workflows. This includes authentication integration, the Flutter lecture player with resume and progress tracking, and the live Q\/A interface for course-specific tutoring. I will implement the lecture video pipeline (slide generation, free TTS voiceover, and avatar lip-sync composition with a reliable fallback). Additionally, I will contribute to analytics presentation and explainability visualization in both teacher and student views to ensure decision transparency. My responsibilities focus on usability, smooth lecture delivery, and ensuring the end-to-end hybrid flow is demonstrable with real data.$$\nI will implement the backend architecture and core AI pipeline. This includes RAG knowledge base creation (extraction, chunking, embeddings, vector retrieval) and lecture\/narration generation from retrieved content. I will also implement assessment screens (quiz\/exam attempt flow), results visualization, and student dashboard views (mastery, confidence trend, recommendations). I will also implement attendance and engagement signal processing, MCQ generation and auto-grading, student model updates, and teacher dashboard APIs. My focus will be on demonstrable technical contribution, stable APIs, database integration, and end-to-end system integration and deployment.$$\n$$\n$$\n$$\n$$\n","comments":"","isDraft":1,"status":1,"created_at":"2026-03-04 03:02:04","updated_at":"2026-03-25 23:23:05","isReviewDraft":1,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1069,"project_id":1495,"title":"ForeSee Health","prob":"The rapid increase in non-communicable diseases such as diabetes, hypertension, high cholesterol, obesity, and abnormal weight fluctuations has become a major public health concern. These conditions are influenced by poor dietary habits, sedentary lifestyles, and lack of proper health monitoring.\r\n\r\nMany individuals remain unaware of how body metrics, lifestyle patterns, and regional dietary practices contribute to long-term health risks. Existing health and nutrition applications rely on global dietary databases that ignore local food availability, cultural eating habits, and economic affordability. As a result, many diet plans are impractical for users in developing regions.\r\n\r\nCurrent solutions provide reactive health tracking rather than predictive insights, limiting their ability to warn users about future risks.\r\n\r\nThis project addresses these gaps by developing a localized, predictive, and cost-aware digital health management system integrating machine learning\u2013based risk prediction, personalized nutrition planning, and health monitoring to help users make informed lifestyle decisions and prevent chronic diseases.","description":"The proposed application integrates user health data, machine learning models, and cloud-based infrastructure to deliver personalized health insights and dietary recommendations.\r\nThe system begins by collecting user biometric and demographic information, including age, height, weight, BMI, lifestyle habits, and medical history. This information is securely stored in a Supabase (PostgreSQL) cloud database using Row-Level Security (RLS) to ensure user privacy and controlled access.\r\nOnce the data is normalized and stored, it is processed by the Health Risk Analytics Engine, where a Random Forest machine learning model evaluates the user\u2019s health profile. By analyzing patterns within historical clinical datasets, the model predicts the probability of developing chronic conditions such as diabetes, hypertension, or obesity.\r\nBased on the predicted risk levels, the system activates a Proactive Health Warning System, which continuously monitors health indicators and generates alerts when user patterns deviate from recommended health thresholds.\r\nSimultaneously, the Country-Aware Meal Planning Engine analyzes the user's nutritional requirements and dietary goals. Using a localized food database, the system generates meal recommendations based on regional cuisine, nutritional balance, and affordability, ensuring that suggested diets are both practical and accessible.\r\nThe application also includes a Predictive Weight and Progress Forecaster, which utilizes Linear Regression models to estimate future weight trends based on calorie intake, activity levels, and historical data. This allows users to visualize potential outcomes of their lifestyle choices.\r\nAll user interactions and updates are synchronized in real time through Supabase cloud services, ensuring secure data persistence, cross-device accessibility, and continuous health profile updates.$$\nSecure User Authentication & Profile Management: Secure user registration and login with encrypted credentials, Creation and management of personalized health profiles, Secure storage of sensitive health data with privacy controls, User profile updates including biometric and lifestyle information\r\nAI-Driven Health Risk Analytics Engine: Analyze user biometric and historical health data, Predict potential chronic health risks using Random Forest machine learning models, Generate health risk scores and analytical insights, Support data-driven preventive health recommendations\r\nProactive Health Warning & Alert System: Monitor user health metrics and lifestyle patterns, Detect deviations from recommended health thresholds, Generate preventive alerts and health advice, Notify users about potential long-term health risks\r\nCountry-Aware Meal Planning Engine: Recommend meal plans based on regional food availability and cultural dietary habits, Adapt recommendations according to user health goals and predicted risk levels, Calculate calories, macronutrients, and nutritional values per meal, Ensure meal plans remain economically feasible for users\r\nPredictive Weight & Progress Forecaster: Predict future weight trends using Linear Regression models, Track progress based on calorie intake and physical activity, Provide visual dashboards for monitoring fitness goals and outcomes\r\nReal-Time Data Persistence & Cloud Synchronization: Enable real-time data updates and synchronization across devices, Maintain secure health data storage using Supabase cloud infrastructure, Implement automated backup and restoration of user data, Ensure secure and efficient database communication$$\nDesign and implement frontend UI for login and registration\r\nDevelop profile management interface\r\nPrepare dataset and train the Random Forest prediction model\r\nOptimize model accuracy and performance\r\nDevelop alert rules and health threshold detection algorithms\r\nDevelop and maintain a country-specific food and nutrition database\r\nDesign and train weight forecasting models\r\nConfigure Supabase database and real-time synchronization logic$$\nImplement backend authentication logic\r\nIntegrate Supabase database with encryption and secure access policies\r\nIntegrate the trained model with user input data\r\nGenerate and display prediction results within the application\r\nImplement notification system within the application interface\r\nConnect alerts with real-time user data updates\r\nImplement the meal recommendation algorithm\r\nDesign and integrate the UI for displaying meal suggestions\r\nIntegrate predictions into the user dashboard\r\nImplement visualizations such as graphs and progress indicators$$\n$$\n$$\n$$\n$$\n","comments":"","isDraft":0,"status":2,"created_at":"2026-03-13 17:39:48","updated_at":"2026-03-25 23:25:26","isReviewDraft":1,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1052,"project_id":1440,"title":"An AI-Powered Geospatial System for Property Valuation, Livability Scoring and Anomaly Detection","prob":"Traditional real estate platforms in Pakistan function as simple digital classifieds, suffering from three critical flaws. First, price manipulation is rampant; sellers list subjective, inflated prices that mislead buyers and distort market realities. Second, buyers lack objective, quantitative metrics to evaluate a neighborhood's quality, relying solely on physical visits or biased dealer descriptions. Third, these platforms struggle with fraudulent listings, where dealers post statistically improbable prices or spoof locations to bait users. This FYP solves these issues by shifting from a basic \"listing marketplace\" to a \"Geospatial Decision Support System.\" It introduces an AI-powered valuation engine to predict fair market prices, uses spatial algorithms to mathematically score location livability based on nearby amenities, and employs statistical anomaly detection to identify and flag fraudulent listings. This empowers both buyers and sellers with data-driven insights for secure, intelligent real estate transactions.","description":"Smart Estate is an advanced, cross-platform Geospatial Decision Support System designed to elevate real estate transactions through Artificial Intelligence and spatial computing. The system operates on two integrated layers: a core marketplace foundation and an intelligent analytical engine. Users interact with a map-centric mobile application powered by OpenStreetMap, where dealers can list properties and buyers can explore them interactively. However, every action is governed by AI microservices. When a dealer adds a listing, the \"AI Valuation Engine\" evaluates historical data and property features using a Machine Learning Regressor to predict its actual market value, advising the seller on optimal pricing and warning buyers of overpricing. Simultaneously, the \"Behavioral Anomaly Engine\" analyzes the data for statistical outliers\u2014such as price-to-area mismatches or GPS coordinate spoofing\u2014to automatically flag fake or clickbait listings.\r\n\r\nOnce published, the \"Spatial Livability Engine\" queries OpenStreetMap data to calculate the geodesic distance to essential amenities (schools, hospitals, parks), dynamically assigning the property a quantitative \"Livability Score\" (0-100) and generating investment heatmaps. As buyers browse, a \"Collaborative Recommendation System\" tracks their interactions using Cosine Similarity to learn their hidden preferences, automatically suggesting relevant properties. Finally, dealers access a \"Smart Analytics Dashboard\" that provides reverse lead matching (showing active buyer density in their area) and competitive analysis based on AI-generated metrics. This architecture ensures a secure, data-driven, and highly personalized real estate ecosystem.$$\n1. Core Real Estate Marketplace & Interactive Map:\r\n The foundational cross-platform mobile application (Flutter). It enables role-based registration (Buyers\/Dealers\/Seller). Dealers\/Seller can upload comprehensive property listings with exact geographic coordinates. Buyers can visually browse, search, and interact with these property pins directly on a customized, interactive OpenStreetMap interface.\r\n\r\n2. AI Valuation & Seller Pricing Assistant:\r\n A machine learning module utilizing a Random Forest Regressor trained on historical real estate data. For buyers, it acts as a financial safeguard by predicting fair market value and permanently tagging listings as \"Overpriced\" or \"Undervalued.\" For sellers, it functions as a pricing assistant, analyzing market trends to suggest the optimal listing price that maximizes profit while ensuring a fast sale.\r\n\r\n3. Spatial Livability & Heatmap Engine:\r\n A geospatial analysis module that replaces subjective location descriptions. It executes spatial algorithms on OpenStreetMap data to calculate the distance from a property to key infrastructure (schools, parks, hospitals). It aggregates these metrics to compute a standardized \"Livability Score\" (0-100) and renders dynamic heatmaps highlighting high-demand investment zones.\r\n\r\n4. Behavioral Anomaly & Fraud Detection:\r\n A statistical data forensics module designed to secure the platform from scams. It employs anomaly detection algorithms (e.g., Isolation Forest) to identify logical mismatches in listing data. It automatically flags statistically improbable listings, such as severe price-to-area anomalies (clickbait) or location spoofing (where the dropped pin contradicts the written sector\/address).\r\n\r\n5. Collaborative Recommendation System:\r\n An intelligent matchmaking module that replaces static drop-down filters. By analyzing user interaction history (views, saves, search patterns), it builds a user-item interaction matrix. Using Cosine Similarity, it uncovers latent buyer preferences to automatically recommend visually or financially similar properties that users might not have explicitly searched for.\r\n\r\n6. Smart Dealer\/Seller Dashboard: \r\nA dedicated analytical portal empowering dealers with actionable intelligence. Instead of basic view counts, it offers \"Reverse Lead Matching,\" visualizing the density of active buyers searching within their mapped territories. It also provides comparative analytics, showing dealers how their properties' AI valuations and Livability Scores rank against local market competitors.$$\nI shall develop the Artificial Intelligence and Data Analytics layer, which includes three core modules. First, I will build the \"AI Valuation & Seller Pricing Assistant,\" encompassing data preprocessing and the training of the Random Forest Regression model for accurate price prediction. Second, I will develop the \"Behavioral Anomaly & Fraud Detection\" module, implementing statistical machine learning algorithms to identify price outliers and location spoofing. Third, I will implement the \"Collaborative Recommendation System,\" building the user-interaction matrix and Cosine Similarity logic to personalize buyer property feeds. Additionally, I will develop the backend REST APIs (Python\/API) to integrate these AI services with the frontend.$$\nI shall develop the Spatial Computing, Interface, and Systems layer, comprising three core modules. First, I will develop the \"Core Real Estate Marketplace,\" building the cross-platform application UI (Flutter), OpenStreetMap integration, and standard relational database architecture. Second, I will build the \"Spatial Livability & Heatmap Engine,\" implementing the mathematical algorithms to query nearby amenities and generate dynamic map heatmaps. Third, I will develop the \"Smart Dealer\/Seller Dashboard,\" creating the frontend visualization for reverse lead matching and competitive analytics. I will ensure seamless integration of the backend AI data into the interactive map interface.$$\n$$\n1.Decentralized Real Estate Marketplace.\r\n2.Online Real-Estate House Price prediction & Visualization.$$\nAlgorithmic Price Prediction & Anomaly Detection: Shifting from seller-dictated pricing to ML-driven fair value estimation, combined with statistical anomaly detection to automatically flag fraudulent$$\nSpatial Livability Scoring: Utilizing spatial algorithms and OpenStreetMap data to mathematically quantify neighborhood quality (0-100) based on exact distances to schools, parks, and hospitals.$$\nRecommendation Engine: Replacing static, manual drop-down filters with an AI-driven collaborative filtering system that learns buyer preferences to automatically suggest highly recommended Property.","comments":"","isDraft":0,"status":2,"created_at":"2026-02-19 12:55:21","updated_at":"2026-03-25 23:21:16","isReviewDraft":1,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1087,"project_id":1449,"title":"GenResearch: Intelligent Multi-Agent Platform for Academic Research Assistance","prob":"Academic researchers and students spend significant time manually reviewing large volumes of research papers, extracting key insights, writing literature reviews, formatting citations, and drafting research proposals. This process is time-consuming, repetitive, and inefficient, especially when dealing with multiple research sources.\r\nGenResearch addresses this problem by providing AI-powered research agents that automatically analyze research papers, retrieve relevant knowledge, generate summaries, assist in literature reviews, and help draft structured research content.","description":"GenResearch is an AI-powered multi-agent research platform that assists users in analyzing academic documents and generating structured research outputs. The workflow begins when a user logs in through the web interface and uploads research papers or submits a query. The request passes through the Authentication Module and is routed by the API Gateway. The Document Processing Engine extracts text, performs chunking, and generates semantic embeddings that are stored in the Vector Repository.\r\nWhen a task is requested, the AI Agent Execution Engine activates specialized agents. The RAG Retrieval Engine retrieves relevant document chunks from the vector database and provides context to the language model. The Literature Review Agent analyzes multiple papers to generate structured reviews, the Summarization Agent extracts key findings and contributions, the Citation Agent identifies references and formats them in standard styles, and the Proposal Agent drafts structured research proposals. Generated results are returned to the user, while the Monitoring Framework records system activity and performance.$$\nThe platform will be composed of eight major modules, each responsible for a specific layer of the system.\r\nUser Interaction Interface\r\nProvides a web-based interface where users upload research papers, submit queries, and interact with different AI research agents. It displays generated outputs such as summaries, literature reviews, citations, and research proposals.\r\nAuthentication & Access Control\r\nManages secure user login, registration, and authorization. It ensures that only authenticated users can access system features and protects uploaded research data and platform services.\r\nAPI Gateway & Service Router\r\nActs as the central communication layer that receives requests from the frontend and routes them to appropriate backend services, AI agents, and data modules for processing.\r\nDocument Processing Engine\r\nProcesses uploaded research papers by extracting textual content, performing document chunking, and preparing the data for semantic embedding and indexing.\r\nSemantic Vector Database\r\nStores document embeddings generated from processed research papers, enabling efficient semantic similarity search and retrieval of relevant research content.\r\nAI Agent Workflow Engine\r\nManages execution and coordination of specialized AI agents such as literature review, summarization, citation, and proposal generation agents.\r\nRAG Retrieval Engine\r\nImplements Retrieval-Augmented Generation by retrieving relevant document chunks from the vector database and providing contextual knowledge to the language model.\r\nSystem Monitoring & Logging\r\nTracks system performance, logs agent activities, records user requests, and helps evaluate system efficiency and reliability.$$\nAli will develop the following modules:\r\n1. The User Interaction Interface Layer will be built using React and Next.js to allow users to upload papers and interact with AI agents.\r\n2. The Authentication & Access Control will implement user login and access control using backend APIs in FastAPI and database validation.\r\n3. The AI Agent Workflow Engine will coordinate task-specific agents using LangChain\/LangGraph workflows.\r\n4. The System Monitoring & Logging will track system activity, agent performance, and errors using structured logs and evaluation metrics.$$\nKaleem will develop the following modules:\r\n1. The API Gateway & Service Router will be built using FastAPI to manage communication between frontend, AI agents, and backend services.\r\n2. The Document Processing Engine will extract text from PDFs using PyPDF, perform document chunking, and generate embeddings.\r\n3. The Semantic Vector Database will store these embeddings in ChromaDB\/Qdrant to enable semantic similarity search.\r\n4. The RAG Retrieval Engine will implement the RAG workflow using LangChain, retrieving relevant document chunks to provide contextual input to the language model.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-16 19:23:08","updated_at":"2026-03-16 19:23:08","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1091,"project_id":1484,"title":"MediSense AI: A Multi-Modal Healthcare Platform with Bilingual Clinical Documentation Support","prob":"Pakistan's healthcare system faces critical challenges that prevent timely disease detection and efficient medical care. Limited access to diagnostic specialists leaves rural populations unable to obtain neuroimaging analysis, ECG interpretation, or cognitive assessments available only in major cities. Overburdened physicians spend 40-50% of consultation time on manual documentation rather than patient care, reducing service quality. Language barriers exclude 70% of Urdu-speaking patients from reliable health information, while bilingual medical conversations (code-switching between Urdu and English) cannot be accurately transcribed by existing English-only systems. These gaps result in late-stage disease diagnosis, physician burnout, and inequitable healthcare access. MediSense AI addresses these problems by democratizing AI-powered diagnostic screening (Alzheimer's, cardiac conditions, cognitive decline), providing bilingual health assistance, and automating clinical documentation with support for Pakistan's unique code-switched medical conversations enabling early detection, reducing physician workload by 80-90%, and making quality healthcare accessible across all regions.","description":"MediSense AI is a comprehensive intelligent healthcare platform integrating five AI-powered capabilities to address Pakistan's diagnostic and clinical workflow challenges. The system serves both patients and healthcare providers through two specialized portals.\r\nFor patients, the platform provides accessible screening tools: (1) Alzheimer's Disease Detection analyzes uploaded brain MRI scans using Vision Transformer architecture, classifying cognitive decline into four severity levels with 97%+ accuracy; (2) Voice-Based Cognitive Screening enables 60-second smartphone-based assessments using acoustic biomarkers (MFCCs, prosody, pause patterns) to detect early Alzheimer's risk with 75-80% accuracy; (3) Medical Chatbot delivers bilingual health information, symptom assessment, and doctor recommendations in Urdu and English using Gemini 2.0 Flash with retrieval-augmented generation from curated medical knowledge bases.\r\nFor healthcare providers, the platform offers clinical decision support: (4) Cardiac ECG Classification implements the research-grade 3DFMMecg feature extraction method combined with Random Forest classification, diagnosing six cardiac conditions (Normal, Conduction Disturbance, ST\/T-Changes, Myocardial Infarction, Hypertrophy, additional abnormalities) from 12-lead ECG signals with 92%+ accuracy and full clinical interpretability through SHAP explanations; (5) Ambient Clinical Scribe with Bilingual Support employs automatic language detection (FastText) and adaptive routing to transcribe code-switched Urdu-English medical conversations using Whisper API, generating structured SOAP notes via GPT-4\/Gemini, reducing documentation time by 80-90%.\r\nThe platform's unique contribution is comprehensive bilingual architecture addressing Pakistan's linguistic reality where medical conversations routinely mix Urdu, English, and medical terminology. Built using open-source technologies and cost-effective APIs, the system demonstrates feasibility for resource-constrained environments. All models provide explainable predictions aligned with clinical markers, ensuring physician trust and practical deployability in Pakistani healthcare settings.$$\nModule 1: Alzheimer's Disease Detection from MRI Implements Vision Transformer (ViT) architecture for automated classification of brain MRI scans into four Alzheimer's severity categories: Non-Demented, VerMildDemented, MildDemented, and ModerateDemented. Uses Kaggle Alzheimer's MRI 4-Class dataset (6,400 preprocessed images). The ViT model employs patch-based processing (6\u00d76 pixel patches), 8 transformer layers with 4 attention heads, achieving \u226597% accuracy. Users upload MRI scans; the system preprocesses images, generates predictions with confidence scores, and visualizes attention maps highlighting diagnostically relevant brain regions. Includes clinical recommendations for neurologist consultation when cognitive decline is detected.\r\n\r\nModule 2: Cardiac Condition Classification from ECG Employs state-of-the-art 3DFMMecg parametric model for feature extraction from 12-lead ECG signals, followed by Random Forest classification for six cardiac conditions: NORM, CD (Conduction Disturbance), STTC (ST\/T-Changes), MI (Myocardial Infarction), HYP (Hypertrophy), plus additional abnormalities. Uses PTB-XL database (21,837 ECG records). The 3DFMMecg model decomposes signals into five fundamental waves (P, Q, R, S, T), extracting 316 clinically interpretable features including wave amplitudes, locations, shapes, and widths. Achieves \u226592% accuracy with \u226590% MI sensitivity. Provides SHAP-based explanations showing which clinical markers drove predictions, ensuring physician trust.\r\n\r\nModule 3: Intelligent Medical Chatbot Bilingual conversational AI providing health information, symptom assessment, doctor recommendations, and emergency guidance. Implements Gemini 2.0 Flash with Retrieval-Augmented Generation querying MedQuAD knowledge base (47,457 Q&A pairs) and Pakistani hospital directory. Features automatic language detection for seamless Urdu-English interaction, preserving medical terminology with explanatory translations. Includes safety disclaimers, emergency flagging, and refusal mechanisms for diagnosis\/prescription requests. Target: \u226585% user satisfaction with <3 second response time.\r\n\r\nModule 4: Voice-Based Cognitive Screening Non-invasive smartphone-accessible Alzheimer's screening using acoustic analysis of 60-second \"Cookie Theft\" picture descriptions. Extracts 80-dimensional feature vectors combining MFCCs (13 coefficients plus deltas), prosodic features (speech rate, pauses, pitch, energy), and linguistic markers (word count, filler words). Uses Random Forest classification trained on ADReSS Dataset (156 balanced recordings). Target: 75-80% accuracy with \u226575% sensitivity\/specificity. Captures cognitive biomarkers including increased pauses, reduced fluency, and semantic coherence decline.\r\n\r\nModule 5: Ambient Clinical Scribe with Bilingual Support Real-time clinical documentation using automatic language detection (FastText, \u226595% accuracy) and adaptive transcription for code-switched Urdu-English conversations. Processes 10-second audio chunks, routing segments to appropriate Whisper models: English (\u226595% accuracy), Urdu (\u226585%), or Multilingual with medical term preservation. Implements speaker diarization separating doctor\/patient speech. Generates structured SOAP notes using GPT-4\/Gemini, providing drafts for physician review. Automatically deletes audio post-processing for privacy. Target: 80-90% documentation time reduction.$$\nHassan will develop Module 2 (Cardiac ECG Classification), Module 3 (Medical Chatbot), and Module 4 (Voice-Based Cognitive Screening). For Module 2, he will implement the 3DFMMecg parametric model for ECG signal decomposition, extract 316 clinically interpretable features, train Random Forest classifier on PTB-XL database, and integrate SHAP explainability. For Module 3, he will integrate Gemini API with RAG architecture, implement bilingual query processing using MedQuAD knowledge base and Pakistani hospital directory, and develop the conversational interface. For Module 4, he will implement MFCC and prosodic feature extraction using Librosa, train machine learning classifiers on ADReSS dataset, and build the voice recording interface. These modules leverage Hassan's strengths in signal processing, NLP, and API integration.$$\nHaseem will develop Module 1 (Alzheimer's MRI Detection) and Module 5 (Ambient Clinical Scribe). For Module 1, he will implement Vision Transformer architecture, handle image preprocessing and data augmentation, train the model on Kaggle Alzheimer's dataset, integrate attention visualization showing diagnostically relevant brain regions, and build the MRI upload interface. For Module 5, he will implement the complex bilingual pipeline including FastText language detection, adaptive Whisper transcription routing for code-switched Urdu-English conversations, speaker diarization using Pyannote.audio, GPT-4\/Gemini SOAP note generation with proper prompt engineering, and the doctor review interface with privacy-preserving audio deletion. These modules leverage Haseem's strengths in computer vision, multilingual NLP, and clinical workflow integration.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-18 06:07:54","updated_at":"2026-03-18 09:32:53","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1096,"project_id":1486,"title":"TravEra \u2013 An AI-Based Personalized Travel Planning and Booking System","prob":"Traditional travel platforms provide static booking services with fixed pricing and limited personalization. Users must rely on multiple platforms for planning, booking, and communication, which leads to inefficiency and poor decision-making. These systems lack interaction between users and vendors and do not support budget-based optimization or competitive pricing.\r\n\r\nTravEra solves this problem by introducing a centralized platform where users can define travel requirements and vendors can compete through bidding. The system also provides AI-based recommendations and budget optimization, enabling users to make smarter and cost-effective travel decisions.","description":"TravEra is an AI-based travel planning and booking system designed to improve user decision-making rather than just providing booking services. The system allows users to enter travel requirements such as destination, budget, and duration.\r\n\r\nInstead of fixed booking options, vendors submit customized offers through a bidding system. Users can compare these offers based on price, ratings, and trust score before selecting the best option.\r\n\r\nThe system also includes AI-based budget optimization, which suggests travel plans according to user constraints. A vendor trust scoring system ensures reliability based on past performance.\r\n\r\nThis transforms the system from a traditional booking platform into a dynamic travel marketplace.$$\n(1) User Registration & Authentication, (2) AI-Based Recommendations & Budget Optimization, (3) Vendor Integration & Bidding System, (4) Booking & Payment Integration, (5) Vendor Trust & Rating System, (6) Itinerary Customization, (7) Notifications & Reminders$$\nIn TravEra, I shall develop AI-based recommendation system, implement vendor integration and bidding functionality, handle booking and payment system, manage itinerary planning features, build vendor trust scoring system, and perform backend development with proper database integration.$$\nIn TravEra, I shall develop user authentication and secure login system, design user-friendly UI\/UX interfaces, implement notifications and reminders, create budget optimization interface for users, and manage reviews and ratings to improve overall user experience and system usability.$$\n$$\nOnline Travel & Tour \u2013 Web and App-based system$$\nDynamic Vendor Bidding System\r\nUsers submit travel requirements and vendors compete by offering customized packages, enabling competitive pricing and better choices.$$\nAI-Based Budget Optimization\r\nThe system suggests optimized travel plans based on user budget, duration, and preferences using intelligent logic.$$\nVendor Trust Score System\r\nA scoring mechanism that evaluates vendors based on ratings, booking history, and response time to ensure reliability.","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-31 06:42:45","updated_at":"2026-03-31 06:42:45","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1074,"project_id":1494,"title":"VibeApp: A Vibe-Coding Based Platform for Mobile App Generation","prob":"Mobile application development requires programming knowledge and experience with complex frameworks, which creates difficulties for non-technical users. Existing low-code or no-code tools mostly rely on fixed templates and still require manual configuration.\r\nThis project aims to solve this problem by generating basic CRUD (Create, Read, Update, Delete) mobile applications from a natural language prompt. The system will interpret the user\u2019s description and automatically generate a simple mobile app. The generated application will run in a mobile emulator\/simulator, demonstrating the concept of AI-assisted mobile app generation in a controlled environment.","description":"This project proposes a web-based platform that generates basic CRUD mobile applications using natural language input. The user provides a prompt describing the application. The system processes the prompt and uses a machine learning model to convert it into a structured application design. Based on this design, the platform automatically generates Flutter mobile application code. The generated application is then built and executed in a mobile emulator\/simulator, allowing the user to interact with the app and test its CRUD functionality.$$\nIn VibeApp, the modules are: \r\ni. User Interface Module: Allows users to enter prompts and view the generated application.\r\nii. Prompt Processing Module: Processes the user prompt and extracts key information about the required CRUD app.\r\niii. Model Training & Inference Module: Trains a machine learning model using a custom dataset and uses the trained model to interpret prompts and generate structured app specifications.\r\niv. Code Generation Module: Automatically generates Flutter code for the CRUD mobile application.\r\nv. Emulator\/Simulator Module: Runs the generated mobile application in an emulator for testing and interaction.\r\nvi. Backend API Module: Manages communication between the frontend, model, and code generation system.\r\nvii. Storage Module: Stores the training dataset, trained model, generated applications, and related files.$$\nI shall develop the Frontend and Code Generation modules which include:\r\n(i) designing the web-based prompt input interface and user dashboard,\r\n(ii) implementing the prompt processing and structured input formatting system,\r\n(iii) developing the automated Flutter CRUD application code generation system including screens, models, and database operations, and\r\n(iv) handling storage and management of generated projects, datasets, and related files.$$\nI shall develop the Backend and Machine Learning modules which include:\r\n(i) designing and implementing the backend APIs and system orchestration services,\r\n(ii) creating and managing the dataset for training the machine learning model,\r\n(iii) training the machine learning model\r\n(iv) implementing the mobile emulator\/simulator pipeline to run and test the generated mobile applications.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-14 13:12:40","updated_at":"2026-03-14 13:12:40","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1094,"project_id":1506,"title":"Decentralized AI Multi-Asset Marketplace (Services + Real Estate)","prob":"Lack of trust and transparency in freelancing and real estate transactions\r\nPayment fraud, delayed payments, and dependency on middlemen\r\nHigh commission fees charged by intermediaries (agents, platforms)\r\nDifficulty for small investors to access real estate investment opportunities\r\nFake profiles, scams, and unreliable clients\/freelancers\r\nLack of secure ownership and transaction records","description":"The proposed Final Year Project is a Decentralized AI Multi-Asset Marketplace that integrates freelancing services and real estate trading into a single secure and intelligent platform. The system leverages blockchain technology to eliminate intermediaries and ensure transparent, tamper-proof transactions, while artificial intelligence enhances decision-making, matching, and fraud detection.\r\n\r\nIn this platform, users can register as clients, freelancers, or property owners. Freelancers can offer services, while property owners can list and tokenize their real estate assets into digital shares, allowing fractional ownership and investment. Clients and investors can browse, search, and select services or properties based on their requirements.\r\n\r\nThe blockchain layer, implemented through smart contracts, manages all transactions including payments, escrow services, and ownership transfers. Payments are automatically released when predefined conditions are met, ensuring trust and reducing the risk of fraud. All transactions are securely recorded on the blockchain, providing transparency and immutability.\r\n\r\nArtificial intelligence modules are integrated to improve user experience and platform efficiency. The AI system recommends relevant freelancers and properties to users, ranks freelancers based on performance and skills, and detects fraudulent activities or suspicious behavior. Data science techniques are used to analyze user behavior, market trends, and investment opportunities.\r\n\r\nThe system includes a web application and a mobile app for accessibility, along with a cloud-based backend for scalability and data storage. Overall, this project provides a unified, secure, and intelligent ecosystem that addresses real-world challenges in freelancing and real estate by combining AI, blockchain, and modern web technologies.$$\n1. User Management Module\r\n\r\nThis module handles user registration, authentication, and profile management. Users can sign up as clients, freelancers, or property owners. It includes login\/logout functionality, role-based access control, and profile updates. User verification (basic KYC or identity checks) can also be integrated to reduce fake accounts and improve trust.\r\n\r\n2. Freelancing Services Module\r\n\r\nThis module allows freelancers to create service listings (gigs), define pricing, skills, and portfolios. Clients can browse, search, and hire freelancers based on ratings, reviews, and AI recommendations. It also manages job contracts, milestones, and communication between clients and freelancers.\r\n\r\n3. Real Estate Listing & Tokenization Module\r\n\r\nProperty owners can list real estate assets with detailed descriptions, images, and pricing. This module also enables tokenization of properties, converting real estate into digital shares for fractional ownership. Investors can purchase these shares, making real estate investment more accessible.\r\n\r\n4. Marketplace & Search Module\r\n\r\nThis module provides a unified interface where users can browse both services and real estate assets. It includes advanced search, filtering, and categorization features. Users can search based on price, location, skills, ratings, or investment type.\r\n\r\n5. Blockchain & Smart Contract Module\r\n\r\nThis is the core trust layer of the system. Smart contracts are used to:\r\n\r\nHandle payments securely via escrow\r\nAutomate fund release upon task completion or agreement fulfillment\r\nManage ownership transfers for real estate tokens\r\nRecord all transactions on the blockchain for transparency and immutability\r\nThis module eliminates intermediaries and ensures tamper-proof records.\r\n6. AI Recommendation Module\r\n\r\nThis module uses machine learning algorithms to recommend:\r\n\r\nFreelancers to clients based on skills, ratings, and past behavior\r\nProperties to investors based on preferences and market trends\r\nIt improves user experience by providing personalized suggestions.\r\n7. Fraud Detection Module\r\n\r\nThis AI-based module identifies suspicious activities such as fake profiles, unusual transactions, or abnormal behavior patterns. It uses anomaly detection and classification techniques to flag or block potentially fraudulent users, enhancing platform security.\r\n\r\n8. Payment & Escrow Module\r\n\r\nThis module integrates with blockchain smart contracts to handle secure payments. Funds are held in escrow and only released when predefined conditions are met. It ensures both parties (client and freelancer \/ buyer and seller) are protected from fraud.\r\n\r\n9. Web & Mobile Application Module\r\n\r\nThis module provides the user interface for interacting with the platform. The web and mobile apps allow users to:\r\n\r\nRegister and manage accounts\r\nBrowse services and properties\r\nMake transactions\r\nTrack orders, contracts, and investments\r\nIt ensures accessibility and a smooth user experience across devices.\r\n10. Cloud Backend & Data Management Module\r\n\r\nThis module manages the backend infrastructure, APIs, and databases. It ensures scalability, data storage, and system performance using cloud services. It also handles integration between AI models, blockchain network, and frontend applications.$$\ni will handle Blockchain & Smart Contracts Module along with AI-based Recommendation and Fraud Detection Module, responsible for secure transactions, escrow management, real estate tokenization, fraud detection, and intelligent user recommendations within the platform.$$\nI will be responsible for the Web & Mobile Application Module along with User Management and Marketplace Interface.\r\n\r\nUser Registration & Authentication:\r\nImplementation of signup\/login functionality for different user roles (client, freelancer, and property owner), including secure authentication and profile management.\r\nUser Profile Management:\r\nAllows users to create and update their profiles, including personal details, skills (for freelancers), and property details (for owners).\r\nMarketplace Interface (Web & Mobile):\r\nDevelopment of responsive web and mobile application interfaces where users can browse, search, and interact with freelancers and real estate listings.\r\nService & Property Browsing:\r\nEnables users to view, filter, and select services or properties based on categories, pricing, location, ratings, and preferences.\r\nUser Interaction Features:\r\nIncludes messaging\/chat interface, notifications, and dashboards for tracking activities such as orders, contracts, and investments.\r\nFrontend Integration:\r\nIntegration of frontend applications with backend APIs, blockchain transactions, and AI services to ensure a seamless user experience.$$\n$$\nDecentralized Real Estate Marketplace\r\nDEX (Decentralized Crypto Currencies Exchange Platform)\r\nTransparency and Security in Crowdfunding (Decentralized Application-DAPP)$$\nBlockchain-based Smart Contracts for Secure Transactions & Escrow: Automates payments, ensures trustless agreements, and removes intermediaries by releasing funds$$\nReal Estate Tokenization & Fractional Ownership: Converts physical properties into digital tokens, enabling users to invest in fractional shares and increasing accessibility to real estate investments$$\n","comments":null,"isDraft":0,"status":2,"created_at":"2026-03-29 12:16:11","updated_at":"2026-03-29 13:10:15","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1079,"project_id":1437,"title":"Fixify: Home Services Marketplace with Agentic Admin Automation","prob":"Finding reliable home service providers such as plumbers, electricians, cleaners, or technicians is often difficult and time-consuming. Most people rely on personal contacts or informal recommendations, which may not guarantee trustworthy or skilled professionals. Customers also face problems comparing service providers, understanding service costs, and communicating their requirements before booking. This lack of transparency can lead to poor service quality, pricing disputes, and wasted time.\r\n\r\nOn the service provider side, many skilled workers struggle to reach potential customers because they lack a proper digital platform to showcase their services. Additionally, platform administrators face challenges in verifying provider documents, managing registrations, and handling complaints efficiently.\r\nThe proposed system, Fixify, aims to solve these issues by creating a centralized home services marketplace where customers can easily find verified providers, communicate with them, and book services. The platform will also use agentic AI to assist administrators in automating provider verification and improving platform.","description":"Fixify: Home Services Marketplace with Agentic Admin Automation is a digital platform that connects customers with reliable home service providers through a centralized marketplace. The system allows users to easily find, communicate with, and book professionals for household services. The application supports both English and Urdu languages to make the platform accessible to a wider audience.\r\n\r\nThe platform organizes offerings using service as the main category, followed by sub services. Example categories:\r\n\r\nHome Maintenance \u2013 Plumbing, Electrical work, AC repair, Appliance repair, Carpentry\r\nCleaning Services \u2013 Home cleaning, Sofa cleaning, Carpet cleaning\r\nInstallation Services \u2013 CCTV installation, Internet setup\r\nPersonal Services \u2013 Beauty services, Haircut at home\r\n\r\nCustomers can search services or interact with an AI chatbot that recommends the best service providers based on their needs. Customers can also upload images of the required work and communicate with providers through a real-time chat system to discuss service details.\r\n\r\nService providers register on the platform, upload verification documents, and pay a 1000 PKR registration fee through Stripe payment to activate their account. Customers can also make service payments through Stripe, where the platform deducts a 5% commission per service. The admin manages the platform through a web dashboard where agentic AI assists in document verification, provider registration analysis, and platform analytics, while still allowing manual administrative control when required.$$\n1. User Management Module\r\nThis module manages the registration and profile management of customers and service providers. Users can create accounts, log in, and manage their personal information. Service providers can also create profiles and upload required documents for verification. The system supports both English and Urdu languages to ensure accessibility for a wider range of users.\r\n\r\n2. Service Management Module\r\nThis module organizes the services available on the platform using service as the main category and sub services under each category.\r\nHome Maintenance \u2013 Plumbing, Electrical work, AC repair, Appliance repair, Carpentry\r\nCleaning Services \u2013 Home cleaning, Sofa cleaning, Carpet cleaning\r\nInstallation Services \u2013 CCTV installation, Internet setup\r\nPersonal Services \u2013 Beauty services, Haircut at home\r\nCustomers can browse these services and select the most suitable provider.\r\n\r\n3. Service Booking Module\r\nThis module allows customers to request and book services from registered providers. Customers can select the service, choose a provider, upload images of the problem or required work, and schedule the service.\r\n\r\n4. Payment and Commission Module\r\nThe system integrates Stripe payment for secure transactions. Service providers must pay a 1000 PKR registration fee through Stripe to activate their account. Customers can also pay for services through Stripe, and the platform deducts a 5% commission per service.\r\n\r\n5. AI Chatbot Module\r\nThis module provides an AI-powered chatbot that helps customers find suitable service providers. Users can ask for the best provider for a specific service, and the chatbot recommends providers based on ratings and availability.\r\n\r\n6. Real-Time Chat Module\r\nThis module enables direct communication between customers and service providers. Users can discuss service details, negotiate prices, and share information through a real-time chat system. The admin can also view chats if needed.\r\n\r\n7. Agentic Admin Automation Module\r\nThis module assists the admin in managing the platform using agentic AI. It helps with document verification, provider registration analysis, complaint classification, and generating platform analytics while still allowing manual admin control.$$\nFollowing modules to be developed in this :\r\n1-User Management Module\r\n\r\n2- Service Booking Module\r\n\r\n3- Real-Time Chat Module\r\n\r\n4- UI and UX Design of the App$$\nFollowing modules to be developed in this :\r\n1-Service Management Module\r\n\r\n2- Payment and Commission Module\r\n\r\n3- AI Chatbot Module\r\n\r\n4- Agentic Admin Automation Module$$\n$$\nOnline Home Services$$\nAI Chatbot-Based Service Recommendation\r\nThe proposed system introduces an AI-powered chatbot that allows customers to search for services using natural language instead of manually browsing.$$\nAgentic AI for Automated Administrative Tasks\r\nThe system introduces agentic AI to assist the administrator by automating tasks such as service provider document verification, provider registration etc$$\nExpanded Service Categories and Sub Services\r\nThe previous system supported only three services, while the proposed system introduces four main categories with twelve sub services as mentioned above.","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-15 01:09:39","updated_at":"2026-03-15 01:09:39","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1086,"project_id":1439,"title":"GymKey: Tier-Based Multi-Gym Membership Platform with QR Check-In and AI Fitness Guidance","prob":"In the current fitness industry, a gym membership only allows access to a single gym location. Users who travel frequently or want flexibility must purchase multiple memberships, which is costly and inconvenient. On the other hand, independent gyms lack a centralized system to share memberships while securely verifying users and tracking attendance. Existing systems do not support a unified subscription model that works across multiple gyms in a region or country. Additionally, most gym platforms do not provide personalized fitness guidance, which reduces user engagement. This FYP addresses these issues by introducing a tier-based subscription system where a user purchases one membership and gains access to multiple approved gyms of the same tier within a defined region. Secure QR-based check-in ensures real-time verification, fair usage, and transparent attendance tracking","description":"GymKey is a SaaS-based fitness platform that enables users to access multiple partnered gyms using a single subscription. Unlike traditional systems where a membership is limited to one gym, GymKey introduces a tier-based subscription model. Users purchase a subscription linked to a specific tier (e.g., Basic or Premium), which allows them to access any approved gym of the same tier within a region or country.\r\n\r\nThe system consists of a mobile application for users and a web-based dashboard for administrators and gym partners. Gym owners register their gyms on the platform and remain inactive until approved by the administrator. Once approved, gyms are assigned a tier and can display a dynamically generated QR code at their entrance. Users scan this QR code through the mobile app to check in.\r\n\r\nThe backend verifies the user\u2019s subscription status, tier compatibility, gym approval, and usage rules in real time before allowing entry. Each check-in is recorded for attendance tracking and analytics. The platform also includes an AI-based personalized fitness guidance module that provides users with workout and calorie recommendations based on their fitness profile. GymKey is designed as a scalable and secure SaaS solution suitable for real-world deploymen$$\nThe main modules of the GymKey platform are:\r\n\r\nRole-Based Authentication & Access Control Module\r\nProvides secure login and role-based access for Admin, Gym Partner, and User.\r\n\r\nAdmin & System Management Module\r\nEnables administrators to approve or reject gyms, manage tiers and subscriptions.\r\n\r\nUser Mobile Application Module\r\nAllows users to register, manage profiles, purchase subscriptions, view gyms, and check in using QR codes.\r\n\r\nGym Partner Onboarding & Approval Module\r\nAllows gym owners to apply for partnership and await admin approval before activation.\r\n\r\nGym Partner Dashboard Module\r\nEnables gyms to view attendance, generate QR codes, and monitor daily check-ins.\r\n\r\nTier-Based Subscription Management Module\r\nManages subscription plans linked to gym tiers and regions rather than individual gyms.\r\n\r\nQR-Based Check-In & Attendance Module\r\nVerifies user access in real time using time-based QR codes and records attendance securely.\r\n\r\nAI-Based Personalized Fitness Guidance Module\r\nProvides workout and calorie recommendations based on user fitness data using machine learning.\r\n\r\nNotification & Status Management Module\r\nSends approval, rejection, and system notifications to users and gym partners.\r\n\r\nAdmin Analytics & Reporting Module\r\nProvides insights into attendance, subscriptions, and system performance.$$\nI shall develop the core architecture and backend of the GymKey platform. My responsibilities include role-based authentication, admin dashboard, gym onboarding and approval workflow, tier-based subscription logic, QR-based check-in verification, attendance tracking, and system analytics. I will also design and implement the AI-based personalized fitness guidance module, including data preprocessing and recommendation logic. Additionally, I will handle database design, backend APIs, security mechanisms, and deployment of the system$$\nI shall develop the user-facing components of the GymKey platform. My responsibilities include designing the mobile application interface, gym listing and discovery screens, QR scanning interface, user profile management, subscription display, and notification screens. I will also assist in gym partner dashboard UI, system testing, bug fixing, and preparation of project documentation.$$\n$$\n$$\n$$\n$$\n","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-16 19:19:14","updated_at":"2026-03-16 19:19:14","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}{"id":1058,"project_id":1468,"title":"MedScribe: AI-Powered Automated SOAP Note & Billing Scribe for Healthcare","prob":"US doctors face severe burnout, spending over two hours on clinical documentation for every hour of patient care. This leads to fatigue, delayed billing, and reduced patient interaction. This project aims to automate medical documentation by providing an AI-powered scribe that listens to live doctor-patient consultations, automatically generates structured SOAP (Subjective, Objective, Assessment, Plan) notes, and suggests accurate CPT billing codes. The system eliminates manual typing, ensures HIPAA-Compliant data handling, optimizes clinic revenue, and allows healthcare providers to focus entirely on patient care.","description":"The proposed system is a secure, web-based healthcare SaaS platform. Doctors can log in and initiate a live audio recording during a patient consultation. The system uses advanced Speech-to-Text (STT) models to transcribe the conversation in real-time, filtering out casual small talk. An NLP engine powered by Large Language Models (LLMs) then analyzes the medical terminology and categorizes the information into a standard SOAP note format. Furthermore, the system acts as a billing assistant by mapping the extracted diagnoses and treatments to standard CPT billing codes. Doctors can review, edit, approve, and export these finalized notes directly to their Electronic Health Record (EHR) systems via a secure, privacy-first dashboard.$$\n1.\tUser Management & Authentication Module: Manages secure, role-based access for doctors and clinic administrators.\r\n\r\n2.\tLive Audio Streaming & Processing Module: Captures live audio from the consultation and processes it using Speech-to-Text (STT) models.\r\n\r\n3.\tNLP Engine & SOAP Note Generation Module: Analyzes the raw transcript to extract and structure data into Subjective, Objective, Assessment, and Plan sections.\r\n\r\n4.\tMedical Billing & CPT Code Extraction Module: Evaluates the medical conditions and treatments to automatically recommend corresponding standard billing codes.\r\n\r\n5.\tDoctor Dashboard & Note Review Module: Provides an interface for doctors to review, edit, and approve the AI-generated notes.\r\n\r\n6.\tSecure Export & EHR Integration Module: Allows approved notes to be securely downloaded or formatted for hospital record systems.\r\n\r\n7. Admin, Reporting & Audit Trail Module: Monitors system usage, manages clinic settings, and ensures secure data logging.\r\n\u00b7$$\n1.\tUser Management & Authentication Module\r\n2.\tDoctor Dashboard & Note Review Module\r\n3.\tLive Audio Streaming (Frontend WebSockets)\r\n4.\tAdmin, Reporting & Audit Trail Module$$\n1.\tLive Audio Processing & STT Integration\r\n2.\tNLP Engine & SOAP Note Generation Module\r\n3.\tMedical Billing & CPT Code Extraction Module\r\n4.\tSecure Export & EHR Integration Module$$\n$$\nNo Similar fyps$$\n1.\tAutomated SOAP Note Generation from Live Audio$$\n2.\tIntelligent CPT\/Billing Code Extraction$$\n3.\tPrivacy-First Medical Context Filtering","comments":null,"isDraft":1,"status":1,"created_at":"2026-03-09 05:33:00","updated_at":"2026-03-09 05:33:00","isReviewDraft":0,"isInternalDraft":0,"isExternalDraft":0,"reviewedDate":null,"markedDate":null}
No. Project Title Reg. No. Name Date & Time Evaluation Status
1. SmartFit: A Mobile-Based Intelligent System for Realistic Outfit and Hairstyle Recommendation CIIT/SP23-BSE-013/WAH
CIIT/SP23-BSE-025/WAH
 JAMAL AHMED KHAN
MUHAMMAD SINNAN HUSSAIN
 Done Accepted  
2. Hate Speech & Violence Detection in Broadcast Media CIIT/SP23-BCS-078/WAH
CIIT/SP23-BCS-091/WAH
 UNAIZA
SYEDA AMNA WASTI
 Done Accepted  
3. DesignToCode: AI-Powered XML to WPF XAML Generator with Complete MVVM Architecture CIIT/SP23-BCS-005/WAH
CIIT/SP23-BCS-015/WAH
 MARYAM ARIF
AROONA NOOR
 Done Accepted  
4. SmartStudyInstructor: AI-Powered Hybrid Courses Platform with Adaptive, Personalized, Interactive Lecture Videos, Attendance, Engagement, and Assessments CIIT/SP23-BCS-011/WAH
CIIT/SP23-BCS-093/WAH
 RAJA HARIS KHUDADAD
MUHAMMAD ADAN KHURSHID
 Done Accepted  
5. Voice of the Voiceless: Multi-Modal Communication App for Speech-lmpaired Individuals CIIT/SP23-BCS-017/WAH
CIIT/SP23-BCS-043/WAH
 AMNA SAEED
ZAIN NADEEM
 Done Accepted  
6. AI-Driven Fashion Trend Forecaster and Hyper-Personalized Styling Assistant CIIT/SP23-BCS-021/WAH
CIIT/SP23-BSE-010/WAH
 MUHAMMAD USAMA
ZIA-UD-DIN
 Done Accepted  
7. ForeSee Health CIIT/SP23-BCS-019/WAH
CIIT/SP23-BCS-042/WAH
 ALVEENA KHAN
HASHIR HUSSAIN BUTT
 Done Accepted  
8. Aseer: The Shadows of Beloved - FPS Horror Survival Videogame CIIT/SP23-BCS-051/WAH
CIIT/SP23-BCS-111/WAH
 SAAD ALI
AOUN MUHAMMAD
 Done Accepted  
9. An AI-Powered Geospatial System for Property Valuation, Livability Scoring and Anomaly Detection CIIT/SP23-BCS-028/WAH
CIIT/SP23-BCS-074/WAH
 MOIEZ AHMAD
MUHAMMAD HAROON HABIB
 Done Accepted  
10. GenResearch: Intelligent Multi-Agent Platform for Academic Research Assistance CIIT/SP23-BSE-031/WAH
CIIT/SP23-BSE-051/WAH
 ALI AHMED
KALEEM-ULLAH ABBASI
 Feb 17, 2026
10:20 AM 		Revision Pending 
11. A Tokenized Rental Property Platform CIIT/SP23-BSE-027/WAH
CIIT/SP23-BSE-048/WAH
 AYESHA TARIQ
WALIA FAISAL
 Feb 17, 2026
11:20 AM 		Revision Pending 
12. MediSense AI: A Multi-Modal Healthcare Platform with Bilingual Clinical Documentation Support CIIT/SP23-BCS-048/WAH
CIIT/SP23-BCS-086/WAH
 HASSAN RAZA MIR
MUHAMMAD HASEEM
 Feb 17, 2026
11:40 AM 		Pending  
13. TravEra – An AI-Based Personalized Travel Planning and Booking System CIIT/FA22-BSE-113/WAH
CIIT/FA22-BSE-114/WAH
 Sayyab Ashraf
Gulfam Hameed
 Feb 17, 2026
12:00 PM 		Revision Pending 
14. Vouch A QR-Authenticated Subscription Engine for Private and Corporate Reward Networks CIIT/SP23-BCS-052/WAH
CIIT/FA22-BCS-107/WAH
 SYED FAIZAN SHAHID
Rija Sattar
 Feb 18, 2026
10:40 AM 		Revision Pending 
15. SkillLoom – Intelligent Textile Workforce Platform CIIT/SP23-BCS-033/WAH
CIIT/SP23-BCS-038/WAH
 MUHAMMAD SAEED
MUHAMMAD SHAHRUKH KHAN
 Feb 18, 2026
11:40 AM 		Revision Pending 
16. SmartDiet: AI-Assisted Smart Diet Planner Mobile Application with Food Recognition and Chatbot Support CIIT/SP23-BCS-050/WAH
CIIT/SP23-BCS-085/WAH
 MUHAMMAD HANZALA ASHRAF
ZAID JABRAN
 Feb 18, 2026
12:00 PM 		Revision Pending 
17. HamKalaam; a Multimodal Real-Time Pakistan Sign Language (PSL) Recognition and Urdu Sentence Generation System. CIIT/SP23-BCS-060/WAH
CIIT/SP23-BCS-067/WAH
 SHAHBAZ HUSSAIN
TALHA AHMAD
 Mar 3, 2026
09:50 AM 		Revision Pending 
18. VibeApp: A Vibe-Coding Based Platform for Mobile App Generation CIIT/SP23-BSE-040/WAH
CIIT/SP23-BSE-046/WAH
 ZAINAB EHSAN
AQSA ARIF
 Mar 3, 2026
10:35 AM 		Revision Pending 
19. EstimaTec: A Web-Based Construction Estimation and Project Workflow Management System CIIT/SP23-BCS-016/WAH
CIIT/SP23-BCS-054/WAH
 MUHAMMAD TAYYAB
MUHAMMAD AWAIS
 Mar 4, 2026
09:50 AM 		Revision Pending 
20. Static and Dynamic Analysis Framework for Detecting Delphi-Based Remote Access Trojans (RATs) CIIT/SP23-BCS-009/WAH
CIIT/SP23-BCS-066/WAH
 TAIMOOR MALIK
ZAINAB SAIF
 Mar 4, 2026
10:15 AM 		Revision Pending 
21. Decentralized AI Multi-Asset Marketplace (Services + Real Estate) CIIT/SP23-BCS-075/WAH
CIIT/SP23-BCS-104/WAH
 ABDULLAH TAHIR
AYESHA ALI
 Mar 4, 2026
10:35 AM 		Pending  
22. Fixify: Home Services Marketplace with Agentic Admin Automation CIIT/SP23-BSE-026/WAH
CIIT/SP23-BSE-039/WAH
 NAVEED UMAR
MUHAMMAD SAMI
 Mar 5, 2026
09:40 AM 		Revision Pending 
23. GymKey: Tier-Based Multi-Gym Membership Platform with QR Check-In and AI Fitness Guidance CIIT/SP23-BSE-045/WAH
CIIT/SP23-BSE-052/WAH
 HANZALA NASEER
MUHAMMAD USMAN
 Mar 5, 2026
10:00 AM 		Revision Pending 
24. MedScribe: AI-Powered Automated SOAP Note & Billing Scribe for Healthcare CIIT/SP23-BCS-008/WAH
CIIT/SP23-BCS-018/WAH
 MUHAMMAD ABDULLAH
MUHAMMAD AHMAD AFTAB
 Mar 5, 2026
10:40 AM 		Revision Pending 
25. Universal Offline Desktop AI Search Assistant CIIT/SP23-BCS-004/WAH
CIIT/SP23-BCS-110/WAH
 ASAD HAIDER
MUHAMMAD TALHA TARIQ
 Mar 16, 2026
09:00 AM 		Revision Pending