Selected Projects
Selected team projects highlighting my work in Machine Learning.
Insurance Risk Model Auditing: Fairness and Interpretability
This project investigates the predictive performance, interpretability, and fairness of machine-learning models for multi-class life-insurance risk classification. Logistic Regression, Random Forest, and XGBoost were trained and evaluated under substantial class imbalance. The best-performing model was subsequently examined using SHAP-based feature attribution, subgroup performance analysis, feature-ablation tests, and controlled input perturbations.
Backtesting Trading Strategies on FAANG Stocks
This project develops and evaluates rule-based trading strategies using historical data from large-cap technology stocks. Trading signals were constructed from the Relative Strength Index, moving-average crossovers, standardized trading volume, volatility measures, candlestick patterns, and cross-asset return information.