Stop AI Agents from Guessing SQL with Semantic Operator on Kubernetes - Manabu McCloskey, AWS

Why watch

This presentation demonstrates Semantic Operator, a Kubernetes-native open source project that runs Apache Ossie semantic models to prevent AI agents from generating incorrect SQL by misunderstanding business metrics. The demo shows how to manage semantic models as Kubernetes resources, enforce governance rules, validate schemas against physical tables, and expose certified metrics to AI agents via MCP and REST for deterministic SQL generation in StarRocks or Trino. Ideal for developers building AI agents, BI applications, or data platforms requiring governed, deterministic queries.