Projects

In Progress · 2024 – Present

Neural Network Chess Engine

Competitive AI chess with a hybrid classical + neural architecture

A competitive chess engine integrating classical search algorithms with neural network-based evaluation. Built a C++ engine with minimax search and alpha-beta pruning for efficient legal move generation, then integrated a PyTorch neural network for board evaluation, combining classical search with learned heuristics. Trained via supervised learning on expert games and self-play RL, using bitboards and Zobrist hashing for speed.

Features

• —Minimax search with alpha-beta pruning for efficient move generation
• —PyTorch neural network for board evaluation
• —Trained via supervised learning and self-play RL
• —Bitboards and Zobrist hashing for fast board representation

Tech Stack