Real-Time Retail Intelligence: Building E-Commerce Recommendations with Lakebase and AI Search on Databricks
Databricks, Friday, October 2nd, 2026
Reference architecture for a real-time e-commerce recommendation engine on Databricks using Zerobus, Model Serving, Lakebase and AI Search.
Databricks presents a reference architecture for a multi-stage recommendation and ranking engine built entirely on its platform.
Clickstream data is ingested in real time with Zerobus, candidates are retrieved with Databricks AI Search, Lakebase serves online features and Model Serving handles low-latency inference.
The design combines precomputed batch recommendations with a real-time scoring path to deliver personalized responses in under 50 ms, and unifying the pipeline removes glue code, speeds iteration and provides end-to-end lineage.