jobs

Agent Post-Training, Computer Use Research

About the TeamThe Agent Post-Training team creates the frontier agents OpenAI ships to the world. We are training the models behind our agents in Codex, ChatGPT, the API, and other frontier products: persistent, proactive intelligence that

Company

OpenAI

Location

San Francisco

Status

Open

Posted

2026-06-26T17:42:39.603+00:00

OpenAI's Agent Post-Training team, on its computer-use effort, hires a researcher to teach models to operate browsers and desktops. The work covers RL and post-training pipelines, graders, reward signals, evals, environments and data, run through major training runs and into Codex and ChatGPT. The posting asks for technical fundamentals in ML or systems, hands-on experience with LLMs, RLHF/RLAIF or agent training, and comfort moving from vague behavioral failures to concrete experiments. It is unusual for spanning research taste, product impact and load-bearing engineering.