dmodel
Member of Technical Staff
$150K–$225K
San Francisco · On-site · Staff · Full-time · Posted Aug 27, 2026
- Education
- Graduate degree in physics or mathematics, or equivalent research experience (preferred).
- Visa sponsorship
- Visa sponsorship is available; the company commits to a genuine effort when making an offer, supported by immigration counsel.
About the job
Build reinforcement learning environments, evaluations, and research methods to study AI agents’ alignment and interpretability capabilities.
What you'll do
- Build reinforcement learning environments and evaluations, and develop graders robust to specification gaming.
- Run experiments to analyze model behavior and failure modes; investigate interpretability techniques and novel alignment methods.
- Present findings in research meetings and contribute to ML papers.
What they ask for
- Strong proficiency in Python and ML frameworks (PyTorch or JAX).
- Familiarity with alignment and/or interpretability literature.
- Ability to generate research ideas and implement others’ ideas.
- Ability to iterate quickly and collaboratively.
Nice to have
- Published or presented AI safety research, especially Mechanistic Interpretability; academic research experience.
- Experience post-training language models; graduate degree in physics or mathematics, or equivalent research experience.
Pay
$150,000-$225,000 USD base salary
Tools
Python, PyTorch, JAX
AI focus
Research AI alignment and interpretability using reinforcement learning environments, evaluations, and experiments.
Model-written summary · the original posting is authoritative