Matt Yi is Co-Founder and Tech VP at VeloDB—the commercial powerhouse founded by the core team behind Apache Doris—and an Apache Doris PMC Member. With over a decade of expertise in database kernel R&D, Matt has spearheaded major version iterations and driven performance breakthroughs for Apache Doris across diverse enterprise scenarios. As a core driver of both VeloDB's technological roadmap and the open-source community, he has been instrumental in evolving Apache Doris into one of the world's most active and influential data platforms.
Why do promising AI Agents fail when moving from prototype to production? The bottleneck is rarely the LLM—it is the underlying database infrastructure. AI Agents behave differently than human users; they read, write, and query data at a scale and frequency that traditional systems simply cannot handle. Join Matt Yi, Co-Founder and Tech VP at VeloDB, as he unravels why production AI Agents need a fundamentally different data engine. This session bypasses high-level theory to deliver hands-on architectural patterns, real-world engineering lessons, and a practical guide on how to architect a single, unified database engine for the next generation of AI.
