Deep Reinforcement Learning Agent

• Led a team in developing a deep reinforcement learning agent for SuperTux Ice Hockey, implementing a Deep Q Network (DQN) algorithm using PyTorch.
• Achieved a 75% win rate against built-in AI opponents, showcasing the effectiveness of the implemented deep learning model.
• Orchestrated an end-to-end machine learning pipeline, including:
  • Data curation and generation
  • Preprocessing of over 1 million game states
  • Model training and optimization
• Authored a comprehensive technical report adhering to academic standards, detailing:
  • Methodology
  • Experimental design
  • Results analysis

Technologies

• Utilized a robust tech stack for implementation and collaboration:
  • Python
  • PyTorch
  • NumPy
  • Pandas
  • Git