A reinforcement learning-powered robotic control system that achieved 90% success rate in object manipulation tasks by implementing DDPG algorithm with PyTorch and integrating real-time computer vision capabilities.

A scalable location-based web application that helps groups find optimal meetup locations by leveraging MongoDB's geospatial queries and Google Places API integration.

An award-winning healthcare solution that analyzes gait patterns using KNN prediction models to predict fall risk, validated through real-time demonstrations with over 20 participants and winning first place in the Innovation4Health Hackathon.

Developed a robust self-checkout system backend using Spring Boot microservices, handling 500+ transactions per minute with real-time inventory updates and distributed payment processing. The system leverages Redis caching and PostgreSQL for high performance, maintaining sub-50ms response times and 98% test coverage.