Researcher, Agentic Post-Training
OpenAI
San Francisco, California$295,000 - $445,000 est.Full-timePosted 12h ago
Team Description OpenAI is looking for exceptional researchers to join the Post-Training Frontiers team, which is responsible for post-training the agentic models we ship across Codex, the API, ChatGPT Thinking, and ChatGPT Pro. The Post-Training Frontiers team sets up the pipeline for deciding which integrations can go into the post-training run, develops its own horizontal improvements to the model, and trains the final model. The role requires working on the most impactful horizontal improve…
The average job posting receives 250 applications.
Stand out by tailoring your resume to this specific role. Our AI resume builder highlights the skills and experience that matter most to this employer.
More open roles at OpenAI
More software engineering roles in San Francisco
Explore related listings
Frequently asked questions
Who is hiring for Researcher, Agentic Post-Training at OpenAI?+
OpenAI is actively hiring for this Researcher, Agentic Post-Training role. Click "Apply Now" to submit your application directly on OpenAI's careers page — Careeronaut doesn't charge employers or candidates for referrals.
When was this Researcher, Agentic Post-Training role posted?+
This listing was first posted on 2026-04-24. We pull the latest copy from the source feed daily, and any role that's taken down gets removed from Careeronaut within seven days so you don't waste time on stale listings.
How should I apply to this Researcher, Agentic Post-Training role?+
Start by tailoring your resume to the posting — most applicants send generic CVs and the first filter recruiters use is keyword relevance. Careeronaut's AI does this automatically: paste the job description, get a matched resume in under a minute, and download as PDF or DOCX.
Where can I find more Software Engineering jobs in San Francisco?+
Browse all open software engineering roles in San Francisco on the listing pages linked below. You can filter by salary, remote-friendly, and posting date.