Software
Viktor Kessler
Karen AtabekjanPPedro SilvaJoin us for the Apache Iceberg™ Europe Meetup! Our event is hosted in Amsterdam co-hosted by booking.com and Vakamo.
🎟️ When registering, please select one of the two ticket types:
5:00 pm – Doors Open + networking
6:00 pm – Awesome talks starts
🎙️Matt Sidley (Snowflake): Inside Iceberg V4
🎙️ Shreyam Sinha (Booking.com): From AWS Glue to Snowflake Horizon: An Iceberg Migration in Two Phases
🎙️ Viktor Kessler (Vakamo): Agentic Readiness Starts in the Iceberg Catalog
19:30 - 19:50 Networking break
🎙️Amit Gilad (lakeOps): Designing a Multi-Engine Lakehouse with Apache Iceberg: One Table, Many Engines
🎙️Will Martin (Dremio/SAP): Delivering Data Sovereignty in the Agentic Era With OSS
20:30 More Networking
21:00 Meetup ends
booking.com
Oosterdokskade 163
1011 DL Amsterdam
🪪 ID Requirements
bring an ID with your
V4 restructures the Iceberg metadata layer rather than adding to it, and it is being driven by workloads the format was not originally designed for: streaming pipelines committing every few seconds, ML feature tables with thousands of columns, and disaster recovery that demands table portability. This talk covers the shape of those changes: how table metadata is being restructured, how tables are becoming less tied to a physical location, how the spec is growing to carry data types that used to live outside the table, and what all of that means for engines that have to read and write it. Drawn from Snowflake's work implementing V4 and contributing proposals upstream.
Matt Sidley is a Senior Manager of Product Management at Snowflake, where he leads product for the data lake and storage platform, including Snowflake's Iceberg storage and interoperability work. He works closely with the Snowflake engineers contributing V4 spec proposals to Apache Iceberg.
Booking.com’s move from Hadoop to AWS enabled a growing Iceberg data lake alongside widespread Snowflake adoption. It also left us coordinating two catalogs, requiring synchronization machinery, different table representations, and split operational responsibilities.
This session examines our ongoing evolution from AWS Glue toward Snowflake Horizon and its Polaris-based Iceberg REST interface. We explain the two-phase approach: first route external engines through Horizon while Glue remains the catalog of record, then convert eligible tables to Snowflake-managed Iceberg while retaining customer-owned S3 storage.
We share lessons from testing the new catalog path, updating existing workflows, and preparing for table conversion. Attendees will learn how to stage an Iceberg catalog migration, catch compatibility issues early, and reduce disruption to users.
Shreyam Sinha is a Senior Software Engineer at Booking.com with seven years of experience building and operating large-scale data platforms. His work spans Hadoop, cloud data lakes, and Apache Iceberg, focusing on making data infrastructure reliable, maintainable, and easier to use.
AI agents get smarter by learning from each other: one agent writes a memory or drafts a skill, and the whole fleet reuses it. That reciprocity is also the risk. A single poisoned email can turn into a “helpful” skill that every agent follows the next day, with no record of who wrote it or who approved it. That’s why most agent projects stall at the CISO.
We will talk how to make agents CISO-ready with the same lakehouse machinery that already governs your data. Lakekeeper, the open-source Apache Iceberg REST catalog, gives every agent its own identity, OpenFGA grants and short-lived vended credentials scoped to its own datasets. Agents can propose skills but hold no key to publish them. A governance agent triages proposals with versioned rules it can’t change, a human approves, and every decision lands in an audit log and an Iceberg decision log.
Live demo included: three agents, one human, and a catalog that keeps saying no.
Viktor Kessler is Co-Founder of Vakamo and the creator of Lakekeeper, an Apache-licensed Iceberg REST Catalog. He’s a big believer in open standards like Apache Iceberg, which he sees as the backbone of today’s modern, composable Data & Analytics systems.
…