Lakebase and Serving Endpoints Implementation in Real-Time Pricing

This breakout session presents a real-world, one of the first in Europe's FSI Industry, production implementation of Databricks Lakebase and Serving Endpoints for real-time pricing in a large-scale insurance environment. We will walk through how we implemented Databricks-native architecture for low-latency data and model access via REST APIs calls. New architecture is a replacement of external Azure PostgreSQL database and intermediate applications layers.Based on experience at one of the largest insurance companies in Poland we will share how Lakebase was stress-tested on high data volumes and is being introduced to production using baby-steps approach. The session includes architecture decisions, performance observations and code-level examples.Attendees will gain the practical guidance on building end-to-end, real-time and data model serving within Databricks along with a clear understanding of how Lakebase fits into architecture.