Automated Credit Risk Predictive Model with Machine Learning

Develop a comprehensive Python-based credit risk prediction system using pandas, scikit-learn, and TensorFlow. Create a machine learning pipeline that ingests historical loan data, preprocesses financial features, trains multiple classification models (logistic regression, random forest, gradient boosting), and generates a probabilistic risk scoring mechanism. The model should include automated feature engineering, handle class imbalance, implement cross-validation, and produce interpretable risk probability outputs with confidence intervals.