Research Scientist (Control)

The role

You will design and test control protocols and monitors for AI coding agents, using empirical research. You will build evaluation frameworks and datasets, fine-tune open-source models into efficient monitors, and feed new agent failure modes into a shared library. The role is part of a small team building the Watcher monitoring tool.

What you will do

• Design and run experiments to test monitor effectiveness across failure modes and agent behaviors
• Build and maintain evaluation frameworks for monitoring capabilities
• Build and maintain high-quality datasets to train and test monitors
• Iterate on monitoring approaches, balancing detection accuracy with computational efficiency
• Contribute newly discovered coding agent failure modes to the shared failure mode library
• Stay current with AI safety, agent failure, and coding security research
• Fine-tune open-source models to create efficient monitors for high-volume production
• Design agentic monitoring systems that investigate logs to find known and novel failure modes
• Incorporate red-teaming findings into monitor design and evaluation

Required qualifications

• 2+ years of empirical research with large language models or AI systems
• Strong experience with AI coding agents, such as extensive use and comparison of frontier coding agents, or designing or developing them
• Experience with LLM-as-a-judge setups or AI monitoring more broadly
• Experience designing and running experiments, analyzing results, and iterating (for example prompting, scaffolding, agent design, fine-tuning, or RL)
• Strong Python programming skills
• Demonstrated ability to work independently on open-ended research problems

Nice to have

• Familiarity with AI safety concepts, particularly agent-related risks
• Familiarity with computer security, such as security testing and secure system design
• Experience fine-tuning language models or working with smaller open-source models
• Previous work building developer tools or monitoring systems
• Publications or contributions to AI safety or ML research
• Experience with production log systems or production log analysis

Benefits

• Market competitive salary and equity
• Flexible work hours and schedule
• Unlimited vacation
• Unlimited sick leave
• Up to 6 months of paid parental leave
• Comprehensive health, dental and vision insurance
• Retirement savings with competitive employer matching (e.g. 401(k) for US employees)
• Lunch, dinner, and snacks provided on workdays
• Paid work trips, including staff retreats, business trips, and relevant conferences
• $1,000 (USD) yearly professional development budget
• Relocation support and visa fees (if applicable)

Before you apply

• Visa sponsorship is offered in the UK and US, but it is not guaranteed for every role or candidate; if an offer is made, the company will work to find a visa route
• Hiring process: screening interview, a 3-hour take-home test, 3 technical interviews, and a final interview with the CEO
• AI tools may be used on the take-home, and the applicant is responsible for the quality of what they submit
• Applicants are encouraged to apply even if they do not meet all the listed characteristics
• A cover letter is neither required nor encouraged

About the company

Apollo Research focuses on risks from loss of control, especially deceptive alignment and scheming in AI models. It works on scheming detection, mitigation, and control and monitoring research, and works with frontier AI companies. It also builds Watcher, a coding agent security product.