AI Red Teamer

Main mission

Attack AI systems to reveal their vulnerabilities before malicious actors do.

5 key responsibilities

• Attack AI systems: jailbreaks, prompt injections, bypasses.
• Design realistic abuse scenarios by type of malicious actor.
• Document vulnerabilities and work on fixes with teams.
• Build automated adversarial test suites.
• Monitor the evolution of attack techniques in the community.

Key skills

Offensive security, deep knowledge of LLM, adversarial creativity, report writing.

What's expected

Find vulnerabilities before attackers do: a credible red team thinks like the adversary.

Career paths

Lead Red Team, ML Security Engineer, Security Researcher.