OpenAI Codex in secure enterprise environments
I work at OpenAI as a Product Manager focused on deploying Codex into some of the most secure enterprises. My first major deployment has been with NVIDIA, where Codex has moved from early use into company-wide access for engineers and researchers.
What I am focused on
NVIDIA-scale rollout
Leading onboarding and enablement patterns for thousands of users, with workflows that span engineering, research, hardware-farm execution, and less technical enterprise use cases.
Secure agent execution
Partnering with security teams on sandboxing, deterministic controls, permissioning, and review systems that let agents work inside sensitive code and infrastructure.
Deterministic plus LLM review
Exploring where conventional controls should be strict and repeatable, and where LLM review can add judgment, context, and coverage without replacing the security model.
Hands-on product development
Using Codex daily for automations, PRs, Rust programming, workflow design, and direct product feedback to keep the deployment grounded in real technical usage.