A few friends and I explored different ways to build a chess engine by combining classic and modern techniques. We tested and compared four approaches: Minimax, Monte Carlo Tree Search, Convolutional Neural Networks, and LSTM models. Each one had its own strengths depending on the phase of the game. Our goal wasn’t just to win games but to understand how each method thinks and makes decisions. In the end, we proposed a hybrid system that blends the best parts of each model for better overall performance.