Application Deadline

We're currently considering applications on a rolling basis. It can take multiple weeks until we respond, even if you are a great fit.

ABOUT THE OPPORTUNITY

We’re looking for Research Scientists and Research Engineers who are excited to work on safety evaluations, the science of scheming, or control/monitoring for frontier models.

YOU WILL HAVE THE OPPORTUNITY TO

• Work with frontier labs like OpenAI, Anthropic, and Google DeepMind, by running pre-deployment evaluations and collaborating closely on mitigations. See e.g. our work on anti-scheming, OpenAI’s o1-preview system card, and Anthropic’s Opus 4 and Sonnet 4 system card.
• Build evaluations for scheming-related properties (such as deceptive reasoning, sabotage, and deception tendencies). See our conceptual work on scheming, e.g. evaluation-based safety cases for scheming or how scheming could arise.
• Work on the "science of scheming," e.g. by studying model organisms or real-world examples of scheming in detail. Our goal is to develop a much better theoretical understanding of why models scheme and which components of training and deployment cause it.
• Work on automating the entire evals pipeline. We aim to automate substantial parts of evals ideation, generation, running, and analysis.
• Design and evaluate AI control protocols. As agents develop longer time horizons, we're shifting more effort to deployment-time monitoring and other control methods.
• Note: We are not hiring for interpretability roles.

KEY REQUIREMENTS

• We don’t require a formal background or industry experience and welcome self-taught candidates.
• Experience in empirical research related to scheming, AI control, and evaluations, plus a scientific mindset: You have designed and executed experiments, can identify alternative explanations, and can test hypotheses to avoid overinterpreting results. This may come from academia, industry, or independent research.
• Track record of excellent scientific writing and communication: You can understand and communicate complex technical concepts and synthesize scientific results into coherent narratives.
• Comprehensive experience in Large Language Model (LLM) steering, plus supporting Data Science and Engineering skills. LLM steering can include: a) prompting b) LM agents & scaffolding c) fluent LLM usage and workflow integration d) supervised fine-tuning e) RL on LLMs
• Software engineering skills: Our stack uses Python. We’re looking for candidates with strong software engineering experience.
• (Bonus) Experience with Inspect, our primary evals framework, since our recent switch.
• Depending on your preferred role and strengths, we may offer either an RS or RE role.

We strongly encourage candidates to apply even if they don’t meet all requirements. Excellent candidates come from many backgrounds.

LOGISTICS

• Start Date: Target of 2–3 months after the first interview
• Time Allocation: Full-time
• Location: London office (shared with LISA). In-person role; rare partial-remote exceptions possible.
• Work Visas: We can sponsor UK visas

BENEFITS

• Salary: 100k–200k GBP (~135k–270k USD)
• Flexible work hours and schedule
• Unlimited vacation
• Unlimited sick leave
• Lunch, dinner, and snacks provided on workdays
• Paid work trips, including retreats, business trips, and relevant conferences
• $1,000/year professional development budget

ABOUT APOLLO RESEARCH

The rapid rise in AI capabilities offers tremendous opportunities, but also presents significant risks.

At Apollo Research, we’re primarily concerned with risks from Loss of Control—risks arising from the model itself rather than human misuse. We focus particularly on deceptive alignment / scheming, where a model appears aligned but is in fact misaligned and capable of evading oversight.

We work on detecting scheming (evaluations), understanding its science (model organisms), and mitigating it (anti-scheming, control). We collaborate closely with multiple frontier AI companies, including testing models pre-deployment.

At Apollo, we value truth-seeking, being goal-oriented, constructive feedback, and being friendly and helpful. More about our culture can be found here.

ABOUT THE TEAM

The current evals team includes: Mikita Balesni, Jérémy Scheurer, Alex Meinke, Rusheb Shah, Bronson Schoen, Andrei Matveiakin, Felix Hofstätter, Axel Højmark, Nix Goldowsky-Dill, Teun van der Weij, and Alex Lloyd.

Marius Hobbhahn manages and advises the team, though members lead individual projects. You’ll mostly work with the evals team, but will sometimes collaborate with the governance team. Our full team is listed here.

Equality Statement

Apollo Research is an Equal Opportunity Employer. We value diversity and provide equal opportunities regardless of age, disability, gender reassignment, marital status, pregnancy, race, religion, sex, or sexual orientation.

How to Apply

Please complete the application form with your CV. A cover letter is optional. Feel free to share links to relevant work samples.

About the Interview Process

The multi-stage process includes:

1. Screening interview
2. Take-home test (~2.5 hours)
3. Three technical interviews
4. Final interview with Marius (CEO)

Technical interviews are directly relevant to on-the-job tasks. There are no LeetCode-style coding interviews.

For preparation, we recommend hands-on LLM evals projects (e.g. from our starter guide), such as building LM agent evaluations in Inspect.

Your Privacy and Fairness in Our Recruitment Process

We use AI-powered tools to assist with tasks such as resume screening, in compliance with internationally recognized AI governance frameworks.

Your data is handled securely and transparently. All resumes are screened by a human, and final hiring decisions are human-made.

For questions or concerns about data processing or fairness, contact info@apolloresearch.ai.

If you’d like this in PDF, Word, LaTeX, or website-ready HTML, I can generate that too.