Cascade #27 – Yash Sheth: Building the Reliability Layer for Enterprise AI
Yash Sheth is the co-founder and COO of Galileo—a platform helping enterprises measure, test, and scale AI applications with confidence.
Yash’s journey started in Mumbai, moved through investment banking at JP Morgan, and shifted into engineering at Google, where he helped build core parts of the conversational AI stack behind products like Google Assistant and Glass. It was there he saw the real challenge: getting powerful models into production without breaking trust.
That’s what led to Galileo.
What began as a focus on data quality has evolved into a full-stack evaluation platform—used by leading enterprise teams to measure performance, prevent regressions, and ship AI features with confidence. Galileo is now 11x faster and 97% cheaper than traditional eval tools—and built from the ground up for production-scale use.
We talked about Yash’s transition from finance to AI, lessons from scaling models inside Google, and why measurement—not modeling—is the missing layer in most AI stacks. We also got into Galileo’s agentic evaluations, the limits of generic metrics, and what it really takes to deploy AI you can trust.
Learn more about Yash and his company below: