About this role
Role Overview
Evaluate and improve the safety, factual accuracy, and alignment of frontier AI models by reviewing model outputs across complex, policy-sensitive, and ambiguous "grey-area" topics. You will apply safety policies, identify unsafe behaviors, and deliver structured feedback that helps shape model behavior used by large numbers of end users.
Key Responsibilities- Evaluate AI-generated responses for safety, factual accuracy, policy compliance, and overall quality.
- Review content in sensitive domains, including misinformation, political persuasion, self-harm, violence, cyber security, biosecurity, and other high-risk or ambiguous topics.
- Apply, test, and refine evaluation rubrics used for RLHF, SFT, and AI safety benchmarking.
- Detect unsafe outputs, hallucinations, reasoning failures, and policy violations.
- Provide structured, reproducible feedback to improve model alignment and safety performance.
- Collaborate with AI researchers, engineers, and safety teams on ongoing evaluation initiatives.
Required
- Bachelor''s degree or higher in Journalism, Communications, Psychology, Sociology, Public Policy, Law, Biology, Chemistry, Computer Science, or a related discipline.
- 5+ years of professional experience in AI safety, Trust and Safety, journalism, public policy, scientific research, security, or a related field.
- Excellent written English, strong critical thinking, and analytical reasoning skills.
- Ability to consistently evaluate nuanced and policy-sensitive scenarios.
Preferred
- Experience with AI safety workflows, RLHF, SFT, Trust and Safety, or AI evaluation programs.
- Familiarity with safety policies, content moderation, or developing evaluation rubrics.
- Prior experience reviewing complex, high-risk, or ambiguous content.
- Location: Remote.
- Employment type: hourly.
- 60 - 70 hourly.
This role is fully remote. Candidates must meet the required qualifications listed above, including education and a minimum of 5 years relevant professional experience. Strong written English and demonstrated ability to assess nuanced, policy-sensitive content are required.