AI Safety Red Teamer

We are looking for seasoned AI Safety Red Teamers to help a leading AI organization uncover weaknesses in cutting‑edge models through adversarial testing. The focus is on probing frontier systems for unsafe behaviour and policy failures across ambiguous, high‑risk topics. Your day‑to‑day work will involve crafting challenging prompts that stress‑test models, spotting jailbreaks, hallucinations and other safety gaps, and assessing robustness in areas such as misinformation, cyber threats, biosecurity, fraud and political content. You will record findings, contribute to safety benchmark reports, and work closely with AI researchers to improve model alignment and resilience. The role requires a bachelor's degree or higher in Computer Science, Cybersecurity, Journalism, Communications, Psychology, Biology, Chemistry, Public Policy or a related field, plus at least five years of professional experience in AI Safety, Red Teaming, Trust & Safety, cybersecurity, investigative journalism, life sciences or a comparable discipline. Strong analytical reasoning, prompt‑design expertise and clear written communication are essential, as is proven experience creating adversarial prompts or evaluating frontier AI systems. Preferred assets include familiarity with RLHF, SFT, AI Alignment, jailbreak testing, prompt engineering and deep knowledge of grey‑area domains such as cyber, biosecurity, political content or misinformation. What the role asks for: - Bachelor's degree in a listed discipline - 5+ years experience in AI Safety, Red Teaming, Trust & Safety, cybersecurity, investigative journalism or life sciences - Strong analytical reasoning, prompt design and written communication skills - Proven experience designing adversarial prompts or evaluating frontier AI systems - Experience with AI Red Teaming, RLHF, SFT, AI Alignment or Trust & Safety (nice-to-have) - Familiarity with jailbreak testing, prompt engineering or adversarial evaluation methods (nice-to-have) - Expertise in grey‑area domains such as cyber, biosecurity, political content or misinformation (nice-to-have)