Machine Learning Feature Store - Pipeline

Design a feature store for ML systems: feature definition and metadata management, online and offline store architecture, feature computation pipelines, point-in-time correctness guarantees, feature versioning and lineage, monitoring for data drift and feature quality, and integration with ML frameworks (scikit-learn, XGBoost, PyTorch)