AI matching platform to connect donor surplus to children's hospitals in need. Designed in 24 hours, winning at McHacks 2025.

A memory aid device with face recognition, lip reading, and conversation summarization. Designed in 36 hours at Hack The North 2024.

AI mock interview platform that generates behavioral and technical interview questions with instant feedback and performance insights.

A ML trading bot using ensemble learning and time series forecasting with backtesting to identify bullish signals and beat the S&P 500.

Playable ML chess engine trained on grandmaster games to evaluate board states and suggest optimal moves in real time.

Customizable LLM for efficient text generation, fine-tuned using the Tiny Shakespeare dataset to a training loss of 0.32.

A ML model built with RandomForestClassifier for the early detection of diabetes, tuned to an 80% accuracy using the Pima Indians dataset.

A neural network built from scratch with 0 dependencies, trained on the MNIST dataset with a 96% accuracy.