In this workshop, we will build a complete serverless ML system, including a feature pipeline, a training pipeline, an inference pipeline and a UI (Gradio on Hugging Face Spaces) with monitoring. We will use the Hopsworks platform as a data layer for managing your features and models, and Modal as a serverless compute platform for the pipelines. We will introduce the key principles of MLOps and how they are realized in this feature store centric system.
Learn more at Data Innovation's Summit.