Projects
Full-stack AI application that extracts FEN strings from digital chessboard images using a CNN model built from scratch. Features a secure FastAPI backend with async inference and a React + TypeScript frontend for seamless Lichess integration for game analysis.
Multi-agent AI system for automated investment research. Uses CrewAI to orchestrate four specialized agents with parallel execution architecture, significantly reducing inference time. Features an interactive Streamlit dashboard for workflow visualization.
Proof of Proximity (PoP) - Research Thesis
Novel randomized consensus mechanism for IoT networks, optimizing fairness in leader selection while maintaining security and low energy consumption. Features real-time transaction monitoring dashboard and comprehensive performance benchmarking.