Key Responsibilities
Built path-finding for simulations using multi-agent reinforcement learning
Designed Unity simulations in C# to study agent decision-making and coordination
Trained agents for complex tasks and iterated on reward design to fix unwanted behavior
Co-authored reports and presented findings in lab meetings
Notable Achievements
•
Improved agent coordination with reward shaping adapted from video game design principles
•
Produced reproducible demonstrations of coordinated multi-agent behavior in Unity