Simple Flows, Serious Systems: Live Recsys Coding with Metaflow

​Recommender systems don’t have to be complex to be powerful. In this live session, we’ll walk through building an e-commerce recommendation engine with Metaflow - from messy data to deployable product recommendations. The patterns and ML systems topics discussed will generalize across use cases.

​We’ll cover:

• ​The core business problem (cross-selling via user-item interactions)

• ​System design: from preprocessing to modeling to deployment

• ​Why Metaflow’s structure (Flows → Steps → Tasks → Artifacts) keeps things simple, scalable, and collaborative

• ​How we version, experiment, and productionize recommendations without friction

​Expect diagrams, code, and real talk on what it takes to build a practical ML system - live.