Red Team Engineer

Design and execute red team attacks on frontier AI systems, discovering vulnerabilities and delivering actionable security findings.
- Design adversarial scenarios and challenges for Arena competitions and private client engagements.
- Conduct red teaming across prompt injection, tool use, agent exploitation, and multimodal attack surfaces.
- Build internal tools for cheating detection, break verification, and operational support.
- Requires 1+ years AI red teaming, adversarial ML, or security research experience.
- Strong technical writing and independent work across concurrent engagements essential.

Gray Swan is an AI security company that develops tools that automatically assess the risks of AI models.