Job Description
Job DescriptionAbout the Role
Join an early-stage AI safety research team working to develop AI systems that follow laws and legal standards. You will lead research projects from initial design through publication, building on collaboration with legal experts and academic advisors.
What You'll Do
- Design and run experiments to evaluate, train, and control AI agents for legal compliance.
- Build evaluation environments, training pipelines, and runtime harnesses for agent research.
- Develop realistic simulated scenarios from court decisions and legal rules.
- Investigate how agents handle conflicting obligations, overlapping jurisdictions, and legal standards that require judgment.
- Work with legal experts to ground research in practice, then publish findings and release benchmarks and tools.
What We're Looking For
- Two to eight years of experience advancing AI through research, ML engineering, or related work.
- A degree in computer science, machine learning, or a related field, or equivalent research experience.
- Experience with LLM or agent evaluations, training environments, post-training pipelines, or AI safety and alignment.
- Hands-on experience building evaluation harnesses, reinforcement learning environments, or training pipelines.
- Strong Python skills and fluency with the modern ML stack, including production engineering experience.
- A builder-first approach, with experience writing code and shipping research or engineering work.
- Authorization to work in the United States without current or future visa sponsorship.
Compensation & Benefits
Base salary is $185,000 to $210,000 annually. Visa sponsorship is not available.
Location
This is an on-site role in San Francisco, California.