Careers · AI Research
AI Researcher
Original research into how AI systems break, and how to attack and defend the models everyone is rushing to deploy.
Companies are shipping AI faster than anyone can secure it. Your job is to be ahead of the adversary: find the attacks on models and agents before they’re used in anger, prove them, and turn what you learn into protection for our clients and the platform. New ground, real stakes.
What you’ll do
- Research novel attacks on machine-learning and LLM systems: prompt injection, jailbreaks, model extraction, data poisoning, and agent tool-abuse.
- Build proof-of-concept attacks that show real, reproducible impact.
- Turn research into repeatable tests and detections the platform can run for every client.
- Track the field closely and publish what’s safe to share. The best researchers give back.
What you bring
- Strong grounding in machine learning and how modern models and agents actually work.
- A security mindset: you look at a system and see how it fails.
- Hands-on with adversarial ML or LLM security, or the depth to get there fast.
- The ability to read a paper or a codebase and turn it into a working attack.
Bonus points
- Published AI-security research or conference talks.
- Open-source tooling or contributions in the space.
- A background that spans both ML engineering and offensive security.
The setup
Remote or in one of our offices, wherever you do your best work. Independent, which keeps us answering to the work and the people we do it for. We hire slowly and keep people a long time. If that sounds like the place, send your work, not a cover letter.