Cyber Security Engineer - Cyber & Autonomous Systems Team

**Application deadline: 11th October 2026, end of day, anywhere on Earth.** ## About the team The AI Security Institute (AISI) is the UK government's team dedicated to understanding advanced AI risks, working with frontier developers and governments globally. The Cyber & Autonomous Systems team evaluates the capability of frontier and open-weight AI models in cybersecurity, autonomy and AI R&D — using realistic cyber ranges and a large CTF suite, running pre-deployment testing of frontier models, and collaborating with partners across UK government, frontier labs, and the NCSC. ## Role AISI is looking for exceptional cybersecurity experts "at all experience levels, from junior through to senior or staff" who want to work at the forefront of frontier AI security. Day-to-day work includes building state-of-the-art evaluations (cyber ranges), running pre-deployment testing exercises, validating novel model behaviours (models cheating in evaluations), and answering timely research questions (the capabilities of open-weight models). Mentorship and coaching from the manager and technical leads are announced. ## Core responsibilities **Evaluation design & development (60 %)** - Design cyber ranges and CTF-style challenges for automatically grading AI system performance on cybersecurity tasks - Build agentic scaffolding to evaluate frontier models, equipping them with tools such as network packet capture utilities, penetration testing frameworks, and reverse engineering/disassembly tools - Design metrics and interpret results of cyber capability evaluations **Infrastructure engineering (30 %)** - Work alongside other engineers to ensure evaluation environments are robust and scalable **Research & communication (10 %)** - Write reports, research papers and blog posts to share findings with stakeholders - Keep up-to-date with related research taking place in other organisations - Contribute to AISI's broader understanding of AI cyber risks ## Example projects - Onboard and integrate new cyber ranges into the evaluation pipeline - Conduct agent research to improve the cyber capabilities of the team's agents - Improve grading and scoring methodologies for automated evaluation tasks - Integrate defensive telemetry and simulated users into ranges to increase their realism - Collaborate with government partners on joint research publications ## Impact (excerpt — the listing text is cut beyond this point) The work directly shapes the UK government's understanding of AI cyber capabilities and informs safety standards for frontier AI systems.