AI Security Institute

• London
• Director · manages people

Lead the Technical Standards and Best Practice team, setting strategy and overseeing researchers who turn AISI's research into concrete, actionable guidance for frontier AI developers. The role spans technical engagement with developers, evaluators and standards bodies, plus line management of a growing cross-AISI team. It reports to a Deputy Director and is offered at Level 5 or 6, with exceptional candidates at Level 4.

Inspect

• London
• Mid-level

Design and lead rigorous studies on how frontier AI systems influence human decisions, including large-scale RCTs, model evaluations, red teaming and participatory methods. Work spans computational social science, AI safety and machine learning, with opportunities to publish and shape the team's research agenda alongside government and frontier lab stakeholders.

Claude Code · Codex

• London
• Hybrid
• £65,000 – £145,000 pa

The Chem Bio team researches, evaluates and mitigates chemical and biological risks from frontier AI, running capability evaluations, building benchmarks and datasets, and developing safeguards and monitoring methods. The role spans ML, computational biology, chemistry and biosecurity, with findings shared with AI companies and government. It is an expression of interest for research engineers and scientists at any seniority.

• London
• Senior

Runs offensive security engagements against frontier AI evaluation infrastructure, including sandboxes, the open-source Inspect framework, and supporting cloud and identity services. Builds and operates agent-assisted red-teaming harnesses, confirms findings with working exploits, and helps harden cyber ranges with research teams. Findings feed directly into how these systems are built.

AWS · Inspect · Rust +1

• London
• Hybrid · 3 days in office
• Senior

Senior individual contributor building and owning shared platform, tooling and infrastructure for a research team, including LLM-based agent systems, evaluation frameworks and supporting infrastructure. Works closely with interdisciplinary researchers to translate varied needs into maintainable Python systems, with strong engineering discipline around testing, documentation and observability. Also contributes to architectural decisions, engineering standards and mentoring through code review and knowledge sharing.

AWS · Kubernetes · Slurm +2

• London
• Hybrid
• £65,000 – £145,000 pa

Evaluates frontier and open-weight AI models for cybersecurity, autonomy and AI R&D capabilities, building realistic cyber ranges and CTF suites, running pre-deployment testing and validating novel model behaviours. Works alongside engineers and researchers with high autonomy, collaborating with UK government, frontier labs and NCSC.

Cyber Ranges · CTF · Claude