unknown

A behind-the-scenes look at how OpenAI's Codex team operates, based on a conversation with Engineering Lead Thibault Sottiaux. The ~40-person team runs with minimal hierarchy, a single PM, and deep AI integration across every workflow. Codex is used for onboarding new engineers (shipping on day one), automated PR reviews, parallel testing, large-scale refactors, user feedback triage, and planning. The team favors small autonomous squads, bottom-up ideation, and fast iteration. Key prompting practices shared include deep design discussions with the model, critical questioning of implementations, and proactive bug-finding sessions. OpenAI's hiring philosophy favors either strong generalists or deep domain experts who think unconventionally.

How OpenAI's Codex Team Works and Leverages AI

All about Engineering Management, Leadership and Career Growth.