CREDIT RISK - PYTHON SCIKIT-LEARN PANDAS NUMPY

Built a scalable default prediction model (0.876 ROC-AUC), implementing balanced class weights to adapt to severe data imbalances within financial datasets. Engineered complex interaction terms and financial ratios aligned with CFPB and Basel III regulatory frameworks to ensure model explainability and compliance. Designed a dynamic risk-grading system mapping default probabilities to S&P-style tiers (AAA-D), automating underwriting decisions and rate premium recommendations.