AI Safety Red Teamer

This role is for an AI Safety Red Teamer at a company that is working to identify vulnerabilities in advanced AI systems. The company is seeking experienced individuals to test the limits of these systems through adversarial testing. The main day-to-day responsibilities of this role include designing prompts to stress-test AI models, identifying weaknesses and unsafe behaviors, and evaluating model robustness across various sensitive domains. This will involve documenting vulnerabilities and contributing to safety reports, as well as collaborating with AI researchers to improve model alignment and safety. To be considered for this role, candidates must have a Bachelor's degree or higher in a relevant discipline, such as Computer Science or Cybersecurity, and at least 5 years of professional experience in a related field. They must also possess strong analytical and written communication skills, as well as experience designing adversarial prompts or evaluating AI systems. Preferred qualifications include experience with AI Red Teaming, AI Alignment, or Trust & Safety, as well as familiarity with jailbreak testing or prompt engineering methodologies. The company offers the opportunity to work on cutting-edge adversarial testing and influence the development of safer AI systems, alongside leading researchers and safety teams.