Chess Engine
High-performance C# chess engine with alpha-beta pruning, iterative deepening, and an NNUE evaluation model trained on Stockfish Fishtest self-play data.
2026 — 8M+ nodes/sec, depth 18–20
- Developed a high-performance chess engine in C# implementing alpha-beta pruning and iterative deepening, achieving an average search depth of 18 to 20 moves while evaluating 8M+ nodes per second.
- Ensured engine correctness through perft validation, systematically verifying move legality and edge cases.
- Built and trained an NNUE neural evaluation model using PyTorch on 1B+ positions from Stockfish Fishtest self-play games.
- C#
- Python
- PyTorch