Mancala

This project came from an AI fundamentals course challenge: build a playable agent that demonstrates a core search idea in a way people can immediately understand. Our team chose Mancala because each move has a clear, measurable short-term payoff, making it a natural fit for a greedy strategy. The agent evaluates available moves by immediate score impact, then selects the option with the highest direct gain.

I helped translate that simple heuristic into a working game loop and interactive demo. We used Streamlit to keep the interface lightweight so the focus stayed on decision behavior rather than visual complexity. The result was a practical bridge between theory and implementation: a textbook greedy algorithm turned into an opponent that users can play against and inspect move by move.