Control Red Team Research Engineer Research Scientist
UK government research organisation that evaluates advanced AI risks and develops and tests mitigations to inform governments.
About AI Security Institute
The AI Security Institute (AISI) is a research organisation within the UK Department for Science, Innovation and Technology. It conducts technical research, evaluates leading AI systems, develops risk mitigations, and shares evaluation infrastructure such as Inspect.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will design and run machine-learning experiments on control measures, develop adversarial attacks, evaluate monitors and sandboxes, conduct security analyses, and report decision-relevant findings. You will also build reusable experimental tooling, pipelines, and infrastructure for training and serving models.
Requirements
- Ability to design, build, and run machine-learning experiments on frontier models
- Ability to work autonomously on complex research projects involving substantial engineering
- Software engineering and machine learning experience writing clean, documented, reusable code
- Experience with LLM fine-tuning, inference frameworks, or Inspect
- Ability to understand and critique how experiments support safety claims
- Understanding of AI safety and control challenges or willingness to learn quickly
- Collaborative and impact-driven approach
- Experience with reinforcement learning, supervised fine-tuning, evolutionary methods, or similar optimisation methods
- Experience in cybersecurity or security analysis
- Proficient use of LLM coding tools and agents
Responsibilities
- Design and run machine-learning experiments on AI control measures
- Build adversarial attacks that generate evidence about monitor efficacy
- Develop arguments and research outputs on safety claims
- Evaluate frontier-lab monitors and report their implications
- Threat model AI attacker behaviour in frontier deployments
- Test monitors, sandboxes, and surrounding infrastructure
- Conduct security analyses
- Build experimental tooling and pipelines
- Build and operate infrastructure for training and serving models
Benefits
- Pre-release access to frontier models and ample compute
- Hybrid working and flexibility for occasional remote work abroad
- Stipend for work-from-home equipment
- At least 25 days of annual leave
- 8 public holidays
- Extra team-wide breaks
- 3 days of volunteering leave
- Paid parental leave
- Employer pension contribution of 28.97% of base salary
- Cycling, donation, retail, and gym discounts and benefits
Hiring Process
Initial assessment → initial screening call → technical assessment → behavioural interview → research interview → final interview with senior leadership.