Research Scientist, AI Safety
Confirmed on the employer's own hiring board on Nov 3, 2026. First seen by Alion on Oct 31, 2026.
About the company
Our client is an applied research company.
The role
Raydar is recruiting for this role on behalf of our client. Design and run research on safety evaluations for advanced AI models, creating benchmarks and test environments that probe how models behave under pressure and whether they can improve their own training. Work closely with engineers to turn research designs into reproducible evaluations and share findings with outside research partners.
What you'll do
- Set up standardized tests for how well models handle parts of machine learning work and enhance their own training, with defined tasks, baselines and success criteria.
- Build test environments that surface reward gaming, sabotage, deception and failures to respect intended limits.
- Scrutinize evaluations for contamination, weak grading, ambiguous tasks and shortcuts that could distort results.
- Run experiments on current models, review their trajectories and failures, and track how capabilities change over time.
- Partner with engineers to produce reproducible environments with clear scoring and documented limitations.
- Write up results as benchmark releases, research reports and recommendations for partners.
Requirements
What we're looking for
- Hands-on experience with post-training and reinforcement learning.
- Hands-on safety work on large-scale models completed within roughly the past two years.
- Working knowledge of the current capabilities of frontier models and where the field is heading.
- A publication record in peer-reviewed machine learning conferences or established evaluation datasets.
Bonus points
- Experience evaluating models on AI R&D tasks, such as measuring models doing machine learning research or training models.
- Prior work with leading AI labs or government AI safety bodies, including pre-deployment testing.
- A PhD in computer science or a related field, or a Bachelor's or Master's with substantial additional work experience.
Benefits
Compensation and benefits
- Base salary: USD 250,000 to 450,000 per year
- Equity
- Health, dental and vision coverage
- Retirement plan
Location and work model
- San Francisco, CA, United States
- On-site, 5 days a week
- Full-time