Reinforcement Learning Work
One of the clearest practical AI pieces is a snake-playing agent built with a compact Q-network, replay memory, helper functions, and iterative self-play. The project folders include both manual game versions and training versions, showing the progression from environment setup to agent training and model saving.