Collaborative projects at the intersection of Cybersecurity and frontier AI, where experienced cybersecurity professionals join AI security field leaders to build concrete projects that secure transformative AI systems.
Remote-first, with access to coworking hubs in London, Tel Aviv, and San Francisco for those who want a desk and in-person community.
Why This Matters
AI systems have become deeply integrated into global infrastructure and this trend shows no sign of slowing. Society’s resilience will depend on how secure transformative AI is, from development through deployment. While many people are using today’s AI capabilities, far too few cybersecurity professionals are thinking about how rapidly AI will reshape society or working to mitigate the emerging and worrying threat models.
Projects and Expert Advisors
Projects are proposed and guided by field leaders in AI Security - leading researchers and engineers shaping how AI systems are secured.
Our vision is high quality and productive collaborations that produce publishable and impactful work in a short time frame. Cybersecurity professionals bring domain expertise; field leaders bring frontier-AI context. Together, they build proofs-of-concept, benchmarks, or defensive strategies.
Program Overview
Research teams will work on an impactful project proposed and guided by a frontier AI expert advisor, and produce publishable results, open-source prototypes, or technical reports by the end of the program.
Duration
3 months
September - November 2026
Research Teams
Expert advisor +
2-4 cybersecurity professionals + project manager + technical advisor
Time Commitment
Commitment of 8-30 hours
a week per person
Biweekly programming at either 15:00-16:00 UTC on Tuesdays or 18:00-19:00 UTC on Saturdays
Locations
Remote
Tel Aviv
London
San Francisco
Format
Two-stage research sprint
Workshop → Conference paper
Support
Research guidance / Compute / APIs / Community / Conference travel funding
Who Should Join
We’re seeking experienced cybersecurity professionals (5 + years) interested in applying their skills to securing transformative AI systems. Ideal participants bring curiosity, technical depth, research experience and the desire to collaborate with AI researchers on real security problems.
Useful skill sets
-
Cryptography (ZK proofs, verifiable compute)
-
AI/ML implementation or agent security
-
Red-team operations and adversarial testing
-
Secure systems and infrastructure design
-
Building proof-of-concepts or attack simulations
-
For more specifics, check the projects! Each has unique needs which may be the perfect fit for your expertise
Focus Areas
AI Infrastructure and Hardware Security
Hardware-level protections (e.g. GPU secure boot, tamper-secure environments), cluster security and TEEs, research into Security Level 4 and SL5, low-bandwidth ML infrastructure
Technical AI Governance
Location attestation, offline licensing, workload attestation (proof of training), model usage verification
Adversarial & Model Security
Jailbreaks and prompt injections, mechanisms for independent audits, protections against model extraction, threat-modeling, backdoor discovery
AI Control & Containment
Zero-knowledge proofs for bounded behavior, containment paradigms, sandboxes, untrusted monitoring
Cybersecurity Evaluations & Demos
Offense–defense balance in AI-driven cyber ops, stealthy intrusion and persistence, real-world vulnerabilities and monitoring, searches for catastrophic misuse cases in the wild
Why Join
-
Apply your cyber skills to frontier AI: Expand your expertise and get hands-on experience working directly with frontier-model security challenges.
-
Build your research portfolio: Co-author papers suitable for top-tier AI security venues and conferences.
-
Create meaningful impact: Contribute to research on some of the most important AI security questions today.
-
Build your professional visibility: Connect with and be seen within the global AI security community.
-
Be part of a strong network: Get access to Heron’s mentorship and career connections.
Key Dates
June-July 2026
Applications open
Early August 2026
Teams announced
1 September 2026
Projects launched
September-November 2026
Weekly research meetings with experts, project manager and technical advisor
October 2026
Mid-project presentations and milestone submission
November 2026
Final submissions and showcase event.
Guaranteed presentation at AI Security Forum in Tel Aviv.
FAQ
Do I need an AI background?
No - strong cybersecurity or cryptography experience is what we value most.
Is this paid?
No, but all teams receive $1000+ in compute, as well as Claude Code Team Accounts for the duration of the Fellowship. In addition, we provide support for conference attendance (up to $5k per team)
Can I participate while working full-time?
Only if you can dedicate at least 8 hours per week.
What if I’m not matched?
You’ll stay in the Heron network for future projects and Forum extensions.
What is the application process?
The application process consists of the application form, a work trial and a short interview, after which candidates will be matched to projects based on team fit.
Can I receive financial support?
If financial support is a blocker, please apply anyways! We may be able to connect you with external funding or scholarships, although we can't commit ahead of time.