AI Safety Evaluator & Model Alignment Expert

Obsidian; Wien; Remote; EUR 70.000 - 110.000; Vollzeit

Zusammenfassung

Obsidian is seeking experienced AI Safety Practitioners to evaluate frontier models on safety, quality, and alignment across complex topics. You will assess responses, apply safety policies, and provide structured feedback to improve model behavior.

Join a team collaborating with AI researchers and safety engineers to advance evaluation rubrics, identify unsafe outputs, and drive improvements in RLHF, SFT, and benchmarking across high-impact domains.

Qualifikationen

• Bachelor's degree or higher in a relevant field.
• 5+ years of professional experience in AI safety, trust & safety, journalism, public policy, or related fields.
• Excellent written English, critical thinking, and analytical reasoning skills.
• Ability to evaluate nuanced, policy-sensitive scenarios consistently.

Aufgaben

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

Kenntnisse

Written English

Critical thinking

Analytical reasoning

Policy evaluation

Ausbildung

Bachelor's degree or higher (relevant field)

Jobbeschreibung

Obsidian is seeking experienced AI Safety Practitioners to evaluate frontier models on safety, quality, and alignment across complex topics. You will assess responses, apply safety policies, and provide structured feedback to improve model behavior.

Join a team collaborating with AI researchers and safety engineers to advance evaluation rubrics, identify unsafe outputs, and drive improvements in RLHF, SFT, and benchmarking across high-impact domains.