Session

Build an Agent You Can Trust: Hands-On Harness, Traces and Evals

A hands-on session where you build a small AI agent and then make it trustworthy. Bring a laptop; we ship code, not slides. We start with a minimal agent loop against an OpenAI-compatible model.

Then we add the parts that separate a demo from something you would run in production: a structured, replayable trace of every step; a small eval suite with five real cases; and a guardrail that fails loudly instead of silently. The best part is the failure. We deliberately break the agent with a prompt injection and watch the trace tell us exactly where it went wrong. Then we fix it and prove the fix with the evals.

You leave with a working repo you can extend after DevFest, and a repeatable loop for evaluating any agent you build.

Harish Kotra

Developer Relations, Hackathons Specialist & A No-Code Educator

Hyderābād, India

Actions

Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.

Jump to top