Deep Think Chess (AI Assisted Engine)

This project showcases the implementation of an AI neural-network-assisted chess engine.The main components of this engine are the Monte Carlo Tree Search and the Transformer, which guides the MCTS tree.In turn, the MCTS tree trains the Transformer to make its predictions more accurate over the course of the game.Besides the intended featured engine we are targeting, we first developed the engine using Minimax with alpha-beta pruning and bitboard manipulation.For the communication protocol, we are using HTTP/2 gRPC for process-to-process communication.The technology stack includes Rust with the Tokio runtime, Next.js, and Python's PyTorch.