Research Scientist, Manipulation Evals
Apart Research · Remote · Europe · San Francisco
Research instituteremote$130k–$200k
At a glance
Highlights
- Design novel evals of AI manipulation informing EU AI Act enforcement.
- Build benchmarks on emotional dependency and multilingual manipulation risks.
- Co-author public papers and reports used directly by EU AI Office.
What they offer
- Real ownership of research at early-stage funded EU AI safety lab with room to grow
- Travel budget for conferences, EU institutional meetings and coworking across Europe
- Learning and development, equipment and home-office budgets
Ideal candidate
Fast-moving empirical AI safety researcher who wants rigorous evals to shape real regulation. You thrive owning research end-to-end, red-teaming your work, and shipping clear findings fast for policymakers.
Requirements
- Prior hands-on evals experience with strong references
- At least one co-authored paper at a top AI/ML venue (workshops/preprints with impact considered)
- Track record of hands-on safety-relevant empirical work such as evals, benchmarks or robustness
- Strong Python, software engineering and data analysis skills
- Fluent with coding agents like Claude Code/Codex for building evals
- Move quickly with high quality bar, clear theory-of-change and red-teaming instincts
- remote policy
- Fully remote, EU-compatible time zones
- hiring process
- Application form, work sample, reference checks, in-depth interviews with Research Scientists and CEO, possibly short paid work trial, offer.
Lab profile
About Apart Research
Nonprofit AI safety organization running research fellowships and sprints to advance safe AI.
What they build
- Research fellowships
- AI safety sprints
- Manipulation and cyber-threat evaluations
Known for
- AI safety field-building
- AI evaluations for governance
- Remote research training
Key facts
- Nonprofit organization
- Remote-first organization
Posted · Last seen