How to Stabilize a GenAI-First Modern Data LakeHouse - Provisioning 20,000 Ephemeral Data Lakes per Year

Explore strategies for stabilizing a GenAI-first modern data lakehouse in this 32-minute conference talk from the Cloud Native Computing Foundation (CNCF). Learn how LinkedIn tackled challenges in scaling their exabyte-scale data lake while introducing GenAI LLMs, migrating to Iceberg, and starting their object storage journey. Discover approaches to maintain platform stability without compromising innovation, focusing on AI and unified SQL. Gain insights into a low-latency system for auto-building lightweight data lakes on Kubernetes for every code commit and PR, and learn how to scale flow failure insights using OpenTelemetry and the JVM. Understand how these techniques enabled the provisioning of over 20,000 ephemeral data lakes per year, catching 2,100 platform issues in the process.