July 15, 2026 - variant data type, WAP workflows, and branching & tagging
Details
## Apache Iceberg™ Meetup Atlanta
Join us on Wednesday, July 15, 2026, at the AWS Buckhead office!
Connect with fellow enthusiasts, share insights, and dive into the latest developments in the Apache Iceberg™ ecosystem! Whether you're a seasoned pro or new to Apache Iceberg, this meetup is the perfect place to exchange ideas and spark innovation.
Agenda
6:00 PM - 6:30 PM: Networking and Welcome Drinks
6:30 PM - 7:45 PM: Welcome Remarks & Presentations!
7:45 PM - 8:30 PM: Demos and Networking
The event will focus on use cases around and innovations in Apache Iceberg (https://iceberg.apache.org/).
We will discuss topics around Open-Source Data Analytics, Open Table Formats (OTF), software concepts like Transactional Data Lakes or Lakehouse, advancements in AI/ML including generative AI, and many more topics of mutual interest that leverage Apache Iceberg.
## Talk 1: Beyond JSON - The Evolution of Variant Data Types in Modern Analytics
As semi-structured data grows exponentially, the Variant data type has become critical for modern analytics, bridging the gap between schema flexibility and query performance. This talk will discuss how analytical systems such as Apache Doris, ClickHouse, Elasticsearch, Snowflake, and Iceberg implement JSON analytics, each with distinct trade-offs. We'll explore technical innovations from dynamic subcolumns and vertical compaction to inverted indexing, revealing why column limits, schema rigidity, and performance bottlenecks persist across platforms. Finally, we'll discuss the architecture convergence toward lakehouse architectures where native JSON optimization, schema evolution, and computational efficiency unite to redefine analytics on open data lakes.
Kevin Shen is a Product Manager at VeloDB.
## Talk 2: Kastor: A Programmable Control Plane for Data Change Management with Apache Iceberg
Implementing Write-Audit-Publish (WAP) workflows at scale is a critical yet unsolved challenge in modern data engineering. While Apache Iceberg has become a foundational format for large-scale analytics, most organizations still rely on manual processes and brittle scripts to manage schema and data evolution—leaving systems vulnerable to downstream breakages, cost overruns, and compliance gaps.
Kastor introduces an intelligence layer and programmable control plane purpose-built for Apache Iceberg. It provides a native interface where automated agents can plan, review, and safely execute schema and data changes within existing Iceberg workflows. Each change is evaluated in real time for cost, performance impact, fragmentation, and downstream risk—enabling guarded automation with traceability, rollback, and human-in-the-loop approvals. The result is a self-auditing, self-documenting platform that accelerates iteration while maintaining trust and stability across the data lifecycle.
Jammond Ruffin is a data systems architect and founder of Kastor Data, where he builds scalable, secure, and intelligent data infrastructure for modern enterprises. Over the past 15 years, he has worked at the intersection of applied machine learning, data engineering, and platform architecture—delivering impactful solutions for organizations including Google/YouTube, and Reddit
## Talk 3: How Apache Iceberg Branching Transforms Data Management
In Apache Iceberg, branching operates according to a particular taxonomy. Branches are named references to a table’s state, similar to branches in Git. They allow you to isolate changes, experiment safely, and manage multiple versions of a dataset simultaneously. For example, a branch might be used to compare to CLONE in Snowflake or Databricks Delta Lake tables, without producing a metadata copy. This approach allows workloads to complete extremely quickly, even in the case of large tables.
Lester Martin, Trino Developer Advocate at Starburst, is a seasoned developer advocate, trainer, blogger, and data engineer focused on data pipelines & data lake analytics using Trino, Iceberg, Hive, Spark, Flink, Kafka, NiFi, NoSQL databases, and, of course, classical RDBMSs. Lester is a polyglot programmer and comfortable in a variety of Dataframe APIs, including Spark, PyStarburst, Ibis, and pandas.
------------------------------------------------
## Future Talks: OPEN FOR PRESENTERS -- do you have a preso and/or demo to share?
Looking for presentations and/or demonstrations related to Apache Iceberg.
Please email lester.martin@gmail.com with your idea OR find us in the #meetup-atlanta channel in the Apache Iceberg Slack - https://iceberg.apache.org/community/#slack.