Inspiration
Tron
What it does
The final system combines neural net predictions with robust backup heuristics to ensure reliability.
How we built it
We combined a PyTorch neural network, pathfinding logic, and backup heuristics in a Flask API matching the contest template.
Challenges we ran into
Making the agent both smart and robust to every edge case, including grid wraparounds and unexpected collisions, was difficult.
Accomplishments that we're proud of
We built an agent that reliably runs, makes intelligent decisions, and integrates a real neural model for game choices.
What we learned
We learned how to train game agents, handle edge cases, and prepare our project for fair and reproducible submission.
What's next for Case Closed Agent
winning!
Log in or sign up for Devpost to join the conversation.