Talks & panels.
Conference sessions and panel discussions — the enterprise data infrastructure argument delivered live, in front of the people who have to build on it.
-
Provided courtesy of and copyright © 2026, DBTA Alliance.
Top Trends in Data Engineering for 2026
DBTA ·
With Sumeet Kumar Agrawal of Informatica and Oz Katz of lakeFS. Moderated by Stephen Faig.
A DBTA roundtable on the direction of data engineering in 2026. My throughline was that the core mission of real-time data hasn't changed even in an era where the buzzword agentic seems to rule all: you still need to get as much data as possible in front of the model as fast as possible—and you need to do it even when you're slammed with new writes. The AI era has just upped the stakes. Whether you're serving a traditional predictive decisioning platform (e.g., fraud detection, recommender systems) or an agent, low-latency data access—even under heavy ingest volume—is the piece neither survives without. It's also the most difficult thing to achieve with current NoSQL stores and the thing most amenable to new thinking in database systems: we've all seen Cassandra-based systems break down under heavy ingest. But the firehose of new data isn't slowing and classic predictive AI and agentic decisionmaking both fail on slow data.