Frontier AI Security Residency
An 8-week, fully funded research & engineering programme in Cambridge, UK, where residents take on concrete cybersecurity & hardware projects to mitigate the risks of frontier AI. Our mission: back the people securing frontier AI via cybersecurity and hardware-enabled verification, before its capabilities outpace our defenses. WHY THIS RESIDENCY Increasingly capable AI models are being deployed more widely than ever before, and the hardware verification and cyber security of frontier AI has not kept pace. Securing the ever growing attack surface for frontier AI across internal and external deployments is one of the hardest and most consequential open problems in the field. The Frontier AI Security residency (FASR) backs engineers and researchers to tackle concrete problems here: verifying what AI compute is doing, keeping models and systems secure, and building the mechanisms that let claims about a system be checked rather than trusted. Residents will work directly with our expert men
Frontier AI Security Residency ERA + Oxford HAIGL + Heron Building the security & verification backbone for frontier AI An 8-week, fully funded research & engineering programme in Cambridge, UK, where residents take on concrete cybersecurity & hardware projects to mitigate the risks of frontier AI. Apply to the residency → Read the FAQ Dates 12 Oct - 4 Dec 2026 Duration 8 weeks, in person Location Cambridge, UK Funding Fully funded Our mission: back the people securing frontier AI via cybersecurity and hardware-enabled verification, before its capabilities outpace our defenses. Why this r
related reading
- Research Fellowship | Heronheronsec.ai
- Community Privacy Residency (2026)community-privacy.github.io
- Ten AI safety projects I'd like people to work onthirdthing.ai
- CS 153: Frontier Systems | Stanford Universitycs153.stanford.edu
- Irregular - Frontier AI Securityirregular.com
- The End-State Fallacy: Where Is AI Security Headed?endstatefallacy.com
- Pacing the Frontierpacingthefrontier.com
- Frontier safety blueprintcdn.openai.com
- How to pace the US frontierblog.aifutures.org
- Frontier Risk Report (February to March 2026) - METRmetr.org
- Focus areas for The Anthropic Instituteanthropic.com
- Emerging processes for frontier AI safety - GOV.UKgov.uk