AI Safety Practitioner – Frontier Model Evaluation & Alignment

About the role

We're looking for experienced **AI Safety Practitioners** to help evaluate the safety, quality, and alignment of frontier AI models across complex, policy-sensitive, and ambiguous ("grey-area") topics. This is a critical trust-and-safety function: you'll be reviewing AI-generated responses against evolving safety policies, flagging risks, and providing structured feedback that directly shapes how these models behave in production.

This role suits detail-oriented thinkers who are comfortable navigating nuanced ethical, legal, and policy questions, and who can apply rigorous judgment even when the "right answer" isn't obvious.

What you'll do

• Evaluate AI-generated responses for safety, factual accuracy, policy compliance, and overall quality
• Review content spanning sensitive domains including misinformation, political persuasion, self-harm, violence, cybersecurity, and biosecurity
• Apply and help refine evaluation rubrics and safety policies based on observed model behavior
• Identify edge cases, policy gaps, and emerging risk patterns in model outputs
• Provide structured, well-documented feedback that supports model improvement and alignment efforts
• Collaborate with policy and research teams to calibrate judgments on ambiguous or contested cases
• Maintain consistency and rigor across large volumes of nuanced evaluation tasks

Requirements

• Strong ability to reason through ambiguous, policy-sensitive, and ethically complex content
• Excellent written communication skills for documenting evaluation rationale clearly
• Familiarity with AI safety concepts, content moderation, trust & safety, or policy evaluation work
• Comfort engaging with sensitive subject matter (e.g., violence, self-harm, misinformation) in a professional, objective manner
• High attention to detail and consistency when applying structured rubrics
• Self-directed and reliable in a fully remote, asynchronous work environment
• Prior experience in AI red-teaming, content policy, research annotation, or related evaluation work is a plus

Compensation

Hourly rate: **$60–$70/hour**, based on experience and evaluation throughput. Flexible, remote, contract-based engagement with opportunities for ongoing work.