5/13/2026 AI Engineers Tech Talk – May Recap SF

The May edition of the SF AI Engineers Tech Talk brought together a room full of engineers, founders, and builders for an evening of no-fluff technical talks. Three founders got specific about the hard problems they actually ran into and how they solved them. If you missed it, the recordings below are worth your time.

Full Event Recording

All three talks back to back, plus Q&A. Good for a lunch break or a commute if you want the full picture of where applied AI engineering is headed right now.

Individual Talks

• Shahram Anver, Co-Founder & CEO at Cleric

Most AI agents treat every incident like it's the first one they've ever seen. Shahram walked through how Cleric is building an autonomous AI SRE that actually retains what it learns, using operational memory and feedback loops to get better over time. The hard part, as he laid out, is keeping that knowledge useful and accurate as production systems keep changing.

• Ronnie Kwesi Coleman, Founder & CEO at Putt

Ronnie got into the mechanics of using LLM-powered agents to automate compliance workflows in regulated industries, shrinking review cycles from months to days. He was honest about the constraints: multi-jurisdictional rules, hallucination risk, and what it actually takes to build systems where being wrong isn't an option.

• Feifan Zhou, Co-Founder & CEO at Tanagram

Feifan took a different angle. Instead of talking about what engineers say they do with AI tools, he looked at what Claude and Cursor transcripts show they actually do. The patterns that came out around debugging, iteration, and problem-solving gave a grounded, data-backed view of how AI-native development is evolving in practice.

Come to the Next One

Events like this happen regularly, and they are always free. Head to fonzi.ai to catch the next one or browse open roles at AI startups in SF and NYC.