My Projects

Dynamic Pricing System

Designed an end-to-end dynamic pricing system with 63+ engineered features from 50K+ historical orders. Built a full ML pipeline achieving R² = 0.881 with Random Forest, and developed a model ensemble combining demand forecasting, price elasticity modeling, and reinforcement learning (30% / 30% / 40% weighting).

Analysis of Pit Stop Errors on F1 Race Results

Processed and merged two F1 datasets (~9,800 race records) using Python (pandas, NumPy) to engineer pit stop error metrics. Built and compared multiple regression models to evaluate pit stop impact on race outcomes, improving model fit (R²: 0.012 → 0.311).