Remote AI Safety Red Teamer — Adversarial Testing Expert

Mercor is seeking a contractor to join as an AI Safety Red Teamer. The role focuses on stress-testing frontier AI models through adversarial prompts, identifying jailbreaks and unsafe outputs, and evaluating robustness across sensitive domains.

You will document findings and contribute to safety benchmarking, working closely with AI researchers to improve alignment and safety. Required are strong analytical reasoning, prompt design, and written communication skills, plus 5+ years in AI safety,