Kyle Weller

You open the Apache Spark™ UI and watch your executors doing full JVM gymnastics. Heap/off-heap flips, Tungsten memory juggling, GC pauses and shuffle spills everywhere. Dynamic allocation struggles, podTemplate nightmares and memoryOverhead prayers are the daily norm. In this 20-minute lightning talk I’ll show you how to drop a new Kubernetes Operator into your Spark-on-K8s cluster and instantly unlock a high-performance vectorized native execution engine. Zero code changes, no new APIs, no rewrites. You’ll see exactly how the operator reconciles every SparkApplication CR and SparkExecutorPod, replacing the JVM gymnastics with true native performance. Plus real production results: 3x+ faster workloads and 40-60% lower cloud spend. Leave with a clear, actionable playbook to upgrade Spark on Kubernetes in one operator deploy.

For years your open lakehouse has been quietly held back by a monolithic, JVM-bound execution engine that was never designed for the speed, freshness, or hardware reality of 2026.

You already see the value of the lakehouse. But are you truly maximizing its potential across all workloads, from BI to AI? In this session, Onehouse unveils how our open lakehouse architecture unifies your entire stack, enabling true interoperability across formats, catalogs, and engines. From lightning-fast ingestion at scale to cost-efficient processing and multi-catalog sync, Onehouse helps you go beyond trade-offs. Discover how Apache XTable (Incubating) enables cross-table-format compatibility, how OpenEngines puts your data in front of the best engine for the job, and how OneSync keeps data consistent across Snowflake, Athena, Redshift, BigQuery, and more. Meanwhile, our purpose-built lakehouse runtime slashes ingest and ETL costs. Whether you’re delivering BI, scaling AI, or building the next big thing, you need a lakehouse that’s open and powerful. Onehouse opens everything—so your data can power anything.

Are you looking for a way to build Iceberg tables for Snowflake? In this session, we will present a simple, battle-tested way to create your lakehouse architecture. We will dig into the toughest challenges of using Iceberg for low latency at scale and we will demo solutions to achieve minute-level freshness for ingestion, and how to run SQL, Spark and dbt pipelines to prepare your data for Snowflake.

The data lakehouse architecture has made big waves in recent years. But there are so many considerations. Which table formats should you start with? What file formats are the most performant?

With a real-time processing engine like Flink and a transactional storage layer like Hudi, it has never been easier to build end-to-end low-latency data platforms connecting sources like Kafka to data lake storage.