Projects

A fully-featured Learning Management SaaS platform designed for interactive, voice-based tutoring powered by AI. This real-time application integrates modern authentication, subscription management, and voice agents to deliver a seamless, immersive learning experience across all devices. This platform was developed as a complete end-to-end SaaS solution with production scalability, modular code architecture, and a focus on reusability and developer ergonomics.

A modern, full-stack web application for screen recording, video uploads, and AI-generated transcripts. It features secure authentication, scalable video delivery, and a responsive UI.

PrepWise is an AI-based interview platform using Firebase for authentication, Vapi voice assistants, and Google Gemini to generate interviews. It offers AI-driven interviews with real-time feedback, detailed transcripts, and a user-friendly interface.

Developed predictive stock price ML models using historical financial data and sentiment analysis, incorporating advanced feature engineering and machine learning techniques. Implemented and compared regression models to evaluate the impact of sentiment data on predictive accuracy, utilising metrics like MAE and R² . Automated data preprocessing pipeline for stock and sentiment data integration, including web scraping, natural language processing, and time-series alignment. Technologies: Python, Scikit-learn, PyTorch, Selenium, BeautifulSoup, Matplotlib, Git

A modern storage management and file-sharing platform designed for seamless file organization and collaboration. Built with Next.js 15 and the Appwrite Node SDK, it offers a robust, user-friendly experience for managing files efficiently. From intuitive file uploads to advanced sharing and search capabilities, this platform showcases cutting-edge technologies for a superior file management experience

A dynamic, user-friendly single-page web application (SPA) built with Django and Vue, offering features like user authentication, profile management, and hobby-based social matching. The app includes AJAX-powered age filtering, friend requests, and dynamic hobby management. Leveraging TypeScript and Python type hints, it ensures type safety and code quality. Comprehensive testing, including Selenium-based end-to-end tests, ensured reliability, while deployment on EECS OpenShift met all accessibility requirements.

A Java Swing-based block puzzle game inspired by Blockudoku, developed as part of an academic project at Queen Mary University of London. The game combines drag-and-drop mechanics with dynamic visual feedback to create an engaging and interactive user experience. Players aim to strategically place pieces on a grid, score points by completing lines or regions, and manage their palette effectively to avoid a game over.

This project is a comprehensive ride-sharing mobile application, designed to provide seamless and efficient transportation services with an emphasis on usability, security, and performance. The application incorporates a host of features that enhance the user experience while ensuring secure transactions and real-time functionalities.

LiveDocs is a Google Docs clone built with Next.js, Liveblocks, and TailwindCSS. It showcases the developer's expertise in creating real-time collaborative environments with a focus on seamless user experience, responsive design, and robust functionality. The project emphasizes the power of real-time updates, ensuring a smooth and interactive document editing experience.

The Healthcare Patient Management Application is designed to simplify the process of managing patient-doctor interactions. It features user-friendly tools for patients and robust administrative functionality, providing seamless communication and operational efficiency