Taught an AI to master Flappy Bird using deep reinforcement learning in PyTorch. I went beyond a basic implementation to explore advanced techniques for more effective learning. • Core Algorithm: Deep Q-Network (DQN) with Experience Replay. • Advanced Enhancements: Implemented both Double DQN (to reduce value overestimation) and a Dueling DQN architecture (to improve state evaluation). • System Design: Engineered a modular agent with a YAML file for easy hyperparameter tuning and experimentation. • Tech Stack: Python, PyTorch, flappy_bird_gymnasium