Algorithmic Credit Scoring Model Pipeline

Construct a comprehensive SQL-based credit scoring pipeline that integrates multiple data sources including financial history, credit bureau reports, and alternative data signals. Implement machine learning feature engineering techniques using advanced window and ranking functions, with specific requirements to handle non-linear risk assessment, create probabilistic default models, and generate interpretable risk scores compliant with financial regulations.