Frontier AI Security Engineer – Cyber & Autonomy

AISafety; Greater London; Hybrid; GBP 65,000 - 145,000; Full time

Job summary

AI Security Institute in London is seeking a Software Engineer to join the Cyber & Autonomous Systems Team, tackling frontier AI security research and evaluation.

You will develop scalable Python software, build evaluation infrastructures, create analysis tooling, and lead tests from design through deployment, while mentoring juniors. The role collaborates with government partners to shape AI safety standards and respond to real-world cyber challenges.

Qualifications

• Production-quality software experience.
• Deep Python expertise with broad tooling familiarity.
• Experience designing software systems and architecture.
• Interest in improving AI safety.
• Ability to own hard, ambiguous problems.
• Strong communication and mentoring skills.
• Cybersecurity or adversarial-security expertise.
• Experience with AI evaluations and deployments.
• DevOps for production systems.
• Knowledge of AI misuse/threat models.

Responsibilities

• Build evaluation infrastructure at scale.
• Develop analysis tooling for results.
• Lead testing exercises and report findings.
• Deliver engineering work from design to maintenance.
• Write production-quality Python with tests and docs.
• Onboard a cyber range to evaluation infra.
• Build an image pipeline for AMIs.
• Improve sandbox service with Core Technology.
• Prepare data to publish research findings.

Skills

Production software

Python

Software architecture

AI safety

Ownership of hard problems

Clear communication

Cybersecurity

AI evaluations

DevOps

AI misuse awareness

Tools

Proxmox

Job description

AI Security Institute in London is seeking a Software Engineer to join the Cyber & Autonomous Systems Team, tackling frontier AI security research and evaluation.

You will develop scalable Python software, build evaluation infrastructures, create analysis tooling, and lead tests from design through deployment, while mentoring juniors. The role collaborates with government partners to shape AI safety standards and respond to real-world cyber challenges.