​Apache Iceberg™ Europe Meetup - live in Berlin!

​​Join us for the Apache Iceberg™ Europe Meetup in Berlin! Our event is co-hosted by Google Cloud, Clickhouse and Vakamo.

​🎟️ When registering, please select one of the two ticket types:

• ​In-Person Ticket: Join us on-site! Your name will be used to pre-register for venue access.

• ​Remote-Only Ticket: Can’t make it in person? No worries—register to join the live stream, receive event recordings, and stay connected with the community.

​Agenda

​5:00 pm – Registration & Networking
6:00 pm – 1st set of short talks
🎙️ Viktor Kessler - Apache Iceberg - Governance and Business Continuity 

​🎙️ Vinnie Dalpiccol - You've Got Iceberg. Now What?

​🎙️ Celeste Horgan - pg_lake: Unifying transactional and analytical data with Postgres

​7:30 pm – 7:50 pm Networking break

​🎙️ Andrei Tserakhau - Iceberg-go: Who Needs a JVM Anyway?
🎙️ Melvyn Peignon - Breaking Down the Iceberg

​8:30 pm – More Networking
9:00 pm – Event close

​Livestream

​

​Presentations & Speakers

​​​​​​​​​​🌟 Apache Iceberg - Governance and Business Continuity 

​​​Modern data platforms increasingly behave like software systems — but most organizations still lack safe workflows for changing data. Updating a table in production can easily break downstream pipelines, dashboards, or machine learning models.

​​​In this session, we explore how Write-Audit-Publish (WAP) and Iceberg’s branching and tagging capabilities introduce Git-like workflows to the data lakehouse. Instead of writing directly to production tables, teams can validate changes, test transformations, and audit results before publishing them.

​​​We’ll walk through the core concepts behind WAP, branches, and tags in Apache Iceberg, and explain how they enable safer experimentation, reproducibility, and controlled data releases.

​​​To make this concrete, the session includes a live demo showing how to:

• ​​​write data to a staging branch

• ​​​validate and audit changes

• ​​​publish the result to production

• ​​​use tags to create reproducible data snapshots

​​​If you’re interested in bringing software engineering best practices to data management, this session will show how Iceberg makes it possible.

​​​​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.

​​​​​​​​​​🌟 You've Got Iceberg. Now What?

​Getting to Iceberg is one problem. Deciding what your platform should actually become is a different one.

​The foundation works. The migration is done. And that's when the harder question shows up: not what do we build next, but what are we actually building toward?

​This talk is about that question. Specifically: what a mature Iceberg-backed platform looks like beyond the storage layer, and how to sequence getting there with a small team and real delivery pressure. We'll cover three pillars: observability, semantics, and data contracts as a coherent vision and the reasoning behind the order we built them in.

​If you've landed on Iceberg and you're staring at the "now what," this one's for you.

​Vinnie leads the data platform team at Staffbase. He came up through Staff Engineer, has run lean platforms across consulting, logistics, and biotech, and spends most of his time thinking about what makes a data platform worth having in the first place. He likes open standards, distrusts architecture that nobody quite remembers choosing, and is not entirely convinced anyone has their platform fully figured out (himself included).

​​​​​​​​​​🌟 pg_lake: Unifying transactional and analytical data with Postgres

​The common wisdom goes: data is either transactional or analytical, and workloads are OLTP and OLAP-shaped to match. The reality of generating and using data is different, and often we need a bit of both. The pg_lake extension brings OLAP workloads to OLTP Postgres.

​This talk will explore the pg_lake extension, including how to build it, and demos of using it with modern data workloads in object storage like Amazon S3. We'll create simple data pipelines with no ETL and high performance analytics.

​Celeste Horgan is a Sr. OSS Developer Advocate at Snowflake and leads the company's Open Source Programs Office. She is also a member of Postgres Europe's DEI Committee and a regular speaker at data engineering meetups in London and across Europe. She got involved in open source technology through the Kubernetes community in 2020, and currently lives in London.

​​​​​​​​​​🌟 iceberg-go: Who Needs a JVM Anyway?

​Let's be honest — the Iceberg ecosystem has a Java addiction. Need to read a table? Spark. Delete some rows? Flink. Check a snapshot? Somehow you're writing XML again.

​But what if you could do all of that with a single compiled binary, no heap tuning, and zero YAML files? iceberg-go is Apache Iceberg's official Go client, and it recently learned some new tricks: equality deletes, row-level mutations, partitioned writes, and multi-table atomic commits. Basically, it grew up.

​In this talk I'll speed-run what makes Go a surprisingly good fit for Iceberg workloads, then live-demo the good stuff: scanning tables, deleting rows by value, and committing mutations through the REST catalog — all without a JVM in sight. I'll show a real CDC ingestion tool I built on top of iceberg-go, and along the way I'll share the spec ambiguities and undocumented behaviors we discovered the hard way — because nothing teaches you a table format like implementing it twice (or three times).

​Come for the hot takes, stay for the equality deletes.

​Andrei Tserakhau is an Engineering Manager at Databricks in Berlin, leading the Table Format Evolution team. With extensive experience as a Tech Lead and Engineering Manager, he's a passionate Go enthusiast and open-source contributor. He actively contributes to Apache iceberg-go, the official Go client for Apache Iceberg, focusing on bringing write-path and DML capabilities to the Go ecosystem.

​​​​​​​​​​🌟 Breaking Down the Iceberg

​Apache Iceberg is often described as a table format, but supporting Iceberg in a query engine requires a deep integration across multiple layers of the data stack: Parquet and Avro file formats, JSON metadata, the Iceberg specification itself, and increasingly the Iceberg REST catalog. Each of these layers introduces performance trade-offs and operational complexity that are not always visible at the SQL level. In this session, we will break down what “Iceberg support” really means in practice and show how ClickHouse’s built-in monitoring and system tables can be used to investigate Iceberg and Parquet efficiency. We will walk through concrete examples of how to analyze metadata access, file pruning effectiveness, scan amplification, and query execution behavior when querying Iceberg tables, and discuss what these observations imply for engine design and table layout choices.