In this webinar, we introduce the challenge of managing historical feature data for ML systems and the role of Delta Lake in low-cost, high-performance storage and querying of feature data. In particular, we present Delta Lake for Feature Stores - data platforms for storing, managing, and serving feature data. Feature stores are the shared data layer that connects ML Pipelines to make complete ML systems. We will show the Hopsworks Feature Store has recently added support for Delta Lake as well as having existing support for Apache Hudi. We will discuss the tradeoffs for the different table formats and the key strengths and benefits of Delta Lake for different ML system ecosystems. In particular, we will look at how Delta Lake enables Hopsworks to build Python-native ML systems.