Design a Credit Risk Scoring System

Design an ML system that predicts the probability of loan default for credit decisions. The system must produce interpretable scores for regulatory compliance, handle missing data common in credit applications, and ensure fairness across protected demographic groups. Consider feature engineering from financial history, model interpretability requirements (SHAP, LIME), and how to monitor for population drift.