Researcher, Alignment CoT Monitorability
OpenAI
San Francisco
More jobs in United States
From the posting
About the Team The CoT Monitorability team at OpenAI studies whether and when the chain-of-thought of frontier reasoning models is monitorable enough to support scalable oversight. We study how to measure monitorability, which training mechanisms affect monitorability, and speculative methods to improve monitorability. While we mostly focus on CoT monitorability at the moment, we care more...
Read the full posting and apply →Posted by OpenAI on its own hiring page. First seen 25 Oct 2026, last confirmed 29 Oct 2026. Details may have changed: check the original before you apply.