Sergey Zavgorodni

What does it take to run Apache Iceberg at the scale of Affirm, Booking.com, DoorDash and Indeed? Join data leaders from all four companies as they share the architecture decisions, cost wins and hard-won lessons from modernizing their data platforms with Iceberg — including how an open lakehouse foundation is accelerating their AI and ML workloads. Walk away with practical frameworks, a clear view of where these teams are investing next and insights you can apply immediately.

Join us as DoorDash shares their battle-tested blueprint for turning Iceberg interoperability into a repeatable, production-grade lakehouse platform. Get insights from their real-world journey adopting Apache Iceberg as the lakehouse table format. Learn about query performance, catalog latency and data optimization, including VARIANT, deletion vectors, CDF/CDC patterns, operationalization via Snowflake catalog-linked databases and Iceberg table replication. We'll cover the “gotchas” of multi-engine architecture patterns, including feature parity gaps, type/schema mismatches and SQL behavior differences.

DoorDash Data organization actively adopts LakeHouse paradigm. This presentation describes the methodology which allows to migrate the classic Data Warehouse and Data Lake platforms to unified LakeHouse solution.The objective of this effort include Elimination of excessive data movement.Seamless integration and consolidation of the query engine layers, including Snowflake, Databricks, EMR and Trino.Query performance optimization.Abstracting away complexity of underlying storage layers and table formatsStrategic and justified decision on the Unified Metadata catalog used across varios compute platforms