Developed a course review directory for University of Toronto students, enabling detailed feedback on courses and instructors. Migrated from EJS-based server-side rendering to a modern ReactJS client-side architecture for a smoother user experience. Built a scalable backend with Node.js and Express.js, and designed a PostgreSQL database hosted on AWS RDS. Deployed the backend on an AWS EC2 instance to ensure high availability and performance.

Developed a Java application using clean architecture to convert playlists between YouTube and Spotify, implementing an algorithm to map and transfer 80+ songs seamlessly. Designed an intuitive Swing interface with drag-and-drop support, real-time progress indicators, and status messages. Adhered to SOLID principles and achieved 90% unit test coverage using JUnit 5.

Built a maze pathfinding game in Python using graph theory concepts and data structures, incorporating cycles for increased difficulty and spanning trees to determine optimal paths. Implemented breadth-first search for efficient traversal through over 150 nodes. Designed a functional and interactive user interface with Tkinter, integrating gameplay mechanics for a smooth user experience.

Developed a rehabilitation-focused workout generator that delivers evidence-based gym routines in seconds. Built with React frontend and Spring Boot backend, featuring AI-powered chatbot for physiotherapy guidance, Excel-based exercise recommendations, and animated exercise demos. Implemented user authentication, progress logging, and personalized workout plans based on muscle groups to strengthen/exclude. Used Docker for containerization and Supabase PostgreSQL for data storage.