EP 45: Why AI Agents Break in Production: The Missing Harness in Your Data Stack

An independent evaluation of Snowflake's Cortex analyst found that six in ten AI-generated queries were wrong, but they all compiled and run without a single error. Today's guest spent the last six months building the fix, and it just topped the industry benchmark for agentic data engineering.

It goes beyond the title with direct discussion of data, like, it's, including: That's not a model problem.

1. 01An independent evaluation of Snowflake's Cortex analyst found that six in ten AI-generated queries were wrong, but they all compiled and run without a single error
2. 02That's not a model problem
3. 03Today's guest spent the last six months building the fix, and it just topped the industry benchmark for agentic data engineering