Credit Risk Modeling — $16.8M Business Impact Optimization
Nov 2025 - Mar 2026 (5 months)
End-to-end ML system for loan default prediction. Optimized decision thresholds achieving 38.3% cost reduction ($16.8M projected savings). Full pipeline: EDA, feature engineering, XGBoost, SHAP explainability, Optuna tuning, Docker deployment. Tech: Python
• XGBoost
• SHAP
• Optuna
• Docker
• scikit-learn