I built APEX, a Formula 1 prediction project that turns FastF1 timing data into per-season Parquet files and predicts lap times and race winners. I evaluated predictions against simple baselines, used chronological data splits, and calibrated 90% conformal prediction intervals.
On Cryptic, I developed a Bitcoin trend prediction pipeline using Random Forest and XGBoost, with a focus on feature engineering and reproducible evaluation. I also built a Flask backend for a Student Management System with GPA computation and Excel export.

