GitNation
Log in
Kiriti Badam

Kiriti Badam

Kiriti Badam is a Member of Technical Staff at OpenAI, where he builds Codex. Previously, he was a founding ML-infrastructure engineer at Kumo.ai, where he helped take the product from an idea to a scalable multi-tenant platform training hundreds of models each day. Before Kumo, Kiriti built large-scale storage and data-processing infrastructure at Google, Databricks, and Samsung. He also teaches practical courses on designing production agentic applications and harness engineering.
OpenAI, USAkiriti-badamsai-kiriti-badam
From Issue to Verified PR: Sizing the Harness for Coding Agents
AI Coding Summit NYCAI Coding Summit NYC
Upcoming
From Issue to Verified PR: Sizing the Harness for Coding Agents
Getting a coding agent to produce a diff is easy. Deciding when that diff deserves a pull request is the engineering problem.

This talk follows one issue through a real coding-agent workflow. We inspect the trace at the first meaningful divergence, then add only the controls that the observed failure justifies: agent-legible context, a checkable completion contract, independent verification, and action boundaries. The comparison becomes a practical framework for sizing the harness to the task: when instructions and tests are enough, when a separate evaluator or sandbox earns its cost, and when a human checkpoint should remain. Attendees will leave with a repeatable method for turning agent failures into systematic improvements instead of longer prompts and more manual review.
best practices