Frontier AI Security Research Engineer

AI Security Institute; Greater London; Hybrid; GBP 65,000 - 145,000; Full time

Benefits offered by this job

Hybrid working

Central London office access

Job summary

AI Security Institute in London is seeking a Research Engineer/Scientist for the Control Red Team to advance AI safety through rigorous ML experimentation and adversarial testing. You will design studies, implement attacks, and help quantify how well monitors prevent harm while collaborating with government-facing partners.

The role welcomes strong researchers/engineers with frontier-model experience, offers significant compute access and a path to leadership on security‑critical projects in a

Qualifications

• Ability to design, build and run ML experiments autonomously.
• Strong software engineering and ML experience: clean, reusable code for experiments.
• Understanding AI safety claims and control challenges.
• Excellent written communication and collaboration across teams.

Responsibilities

• Design and run ML experiments on frontier models.
• Develop adversarial attacks to test monitors and controls.
• Build pipelines to turn questions into results quickly and reproducibly.
• Collaborate with cross-functional teams and publish findings.
• Assess safety claims and contribute to security analyses.

Skills

ML experiments

Red team

Research & engineering

Frontier models

Education

PhD or MSc in relevant field

Tools

Python

ML frameworks

RL

Job description

AI Security Institute in London is seeking a Research Engineer/Scientist for the Control Red Team to advance AI safety through rigorous ML experimentation and adversarial testing. You will design studies, implement attacks, and help quantify how well monitors prevent harm while collaborating with government-facing partners.

The role welcomes strong researchers/engineers with frontier-model experience, offers significant compute access and a path to leadership on security‑critical projects in a