Machine Learning Credit Default Prediction Pipeline

Construct an end-to-end Python machine learning pipeline using scikit-learn and TensorFlow for predicting corporate credit default probabilities. The model should incorporate multi-dimensional financial features including historical financial statements, market sentiment analysis, macroeconomic indicators, and real-time credit rating changes. Implement advanced ensemble techniques like gradient boosting and neural networks, with automated hyperparameter tuning and model interpretability metrics.