FIAM Asset Management Hackathon

2nd place hackathon project combining ML signals and reinforcement learning for dynamic portfolio construction.

This project was built for the FIAM Asset Management Hackathon, where our team won 2nd place. We developed a dynamic portfolio pipeline with feature engineering, model-based stock selection, and reinforcement learning components for monthly rebalancing under practical trading constraints.

The repository includes data preprocessing, baseline portfolio methods, backtesting utilities, and RL experimentation notebooks.