Senior Alignment Researcher, Chain-of-Thought Monitorability
AI research and deployment company with a frontier-model alignment group, San Francisco
| Employer | AI research and deployment company with a frontier-model alignment group, San Francisco (not named) |
|---|---|
| Location | San Francisco, CA (Bay Area) · Hybrid |
| Type & level | Full-time · Senior |
| Pay | USD 295k–500k + equity Source: pay range stated in the source posting |
| Visa path | O-1A fit |
| Role family | AI Safety & Alignment Researcher |
About the role
This team studies whether the reasoning traces of frontier models can be read well enough to support scalable oversight. Its monitoring systems already run on very large reinforcement-learning training runs to catch misbehavior, and the findings feed back into reward and environment design. The role adds research capacity to measure monitorability and to find training choices that preserve it.
What you'll do
- Design and run empirical studies of how readable and monitorable reasoning traces are across frontier models and training setups.
- Build evaluations that test whether monitors can reliably predict properties of interest, including high-stakes misbehavior.
- Investigate how pre-training, synthetic data, mid-training, post-training and reinforcement learning help or hurt monitorability.
- Turn observations from monitoring into hypotheses, experiments and practical recommendations for real training runs.
- Work with researchers and engineers on model training, alignment evaluations, monitoring and frontier-risk teams.
- Write up results as externally publishable research when they advance the science of alignment.
Requirements
- Strong hands-on experience training, evaluating or debugging large ML models, especially language models.
- Depth in alignment, interpretability, model behavior or empirical ML research.
- Ability to turn ambiguous research questions into measurable experiments and weigh noisy or subtle evidence.
- Comfort moving between research ideas and engineering execution.
- Ability to work independently while collaborating across research and engineering teams.
Nice to have
- Direct experience with chain-of-thought interpretability, which is welcome but not required.
Pay & benefits
USD 295k–500k + equity. The posting states equity is offered and that relocation assistance is available for new employees. It does not list other benefits.
Pay source: pay range stated in the source posting. Market data, not an offer from O1DMatch.
Visa path
Researchers in alignment and interpretability often document O-1A evidence such as published papers, citations, peer review or program committee service, invited talks, and a record of original contributions to the field. High compensation and a critical role on a leading team can also support a case. The posting states no citizenship restriction, and an employer like this or a U.S. agent can file the petition. Approval depends on the individual's evidence and is never assured.
How to get matched
- Take the free eligibility checkAn AI review of your background against the eight O-1 criteria, at check.o1dmatch.com.
- Create your profile with code WORKWITHUSIt unlocks the Starter plan at no cost. Tell us your field and the work you do today.
- Upload your evidenceEach document is sorted under the criterion it supports, and your criteria map shows what is still missing.
- Get matchedWe match strong profiles with U.S. employers hiring for roles like this one. Gigs you confirm with verified companies count toward your itinerary, and at a score of 75 we connect you with an immigration attorney.
This listing describes a real U.S. posting so you can see what employers ask for. O1DMatch does not represent that employer and does not hold this opening; applying matches you with employers hiring for this kind of role. No visa outcome is promised.