Frontier AI Safety Evaluator

Mercor; Roma; Remoto; EUR 70.000 - 100.000; Tempo pieno

Descrizione del lavoro

Mercor is seeking experienced AI Safety Practitioners to evaluate safety, quality, and alignment of frontier AI models across complex, policy-sensitive topics. You will assess AI-generated responses, apply safety policies, and help improve model behavior through structured evaluations and feedback.

The role involves reviewing content on misinformation, political persuasion, self-harm, violence, biosecurity, and other sensitive domains; applying and refining evaluation rubrics for RLHF, SFT, and

Competenze

• Bachelor's degree or higher in a related field.
• 5+ years of professional experience in AI Safety, Trust & Safety, journalism, public policy, or related field.
• Excellent written English and analytical reasoning.
• Ability to evaluate nuanced and policy-sensitive scenarios.

Mansioni

• Evaluate AI-generated responses for safety, factual accuracy, policy compliance, and quality.
• Review content involving misinformation, political persuasion, self-harm, violence, and other sensitive domains.
• Apply and refine evaluation rubrics for RLHF and SFT.
• Identify unsafe outputs, hallucinations, and policy violations.
• Provide structured feedback to improve model alignment and safety performance.
• Collaborate with AI researchers and safety teams on evaluation initiatives.

Conoscenze

Excellent written English

Critical thinking

Analytical reasoning

Nuanced content evaluation

Formazione

Bachelor's degree or higher in a related field

Descrizione del lavoro

Mercor is seeking experienced AI Safety Practitioners to evaluate safety, quality, and alignment of frontier AI models across complex, policy-sensitive topics. You will assess AI-generated responses, apply safety policies, and help improve model behavior through structured evaluations and feedback.

The role involves reviewing content on misinformation, political persuasion, self-harm, violence, biosecurity, and other sensitive domains; applying and refining evaluation rubrics for RLHF, SFT, and