This project delivered an effective predictive scorecard that enables lenders to segment loan applicants by risk and make informed decisions. The novel hybrid GAM-Decision Tree model balances accuracy with interpretability, allowing lenders to reliably assess default likelihood. Although the scorecard accurately captures high and low-risk groups, further calibration of the medium-risk group will strengthen model performance in future work. Overall, this scorecard equips lenders with a powerful tool for risk management and supports a more stable mortgage market.