Nuclear Safeguards Strategist for AI Risk Red-Teaming

Mercor; New York (NY); Remote; USD 110,000 - 170,000; Full time

Job summary

Mercor, based in New York, is assembling a panel of nuclear materials and safeguards experts to red-team frontier AI models. The aim is to test whether a model can judge misuse potential while answering legitimate questions fully and refusing genuinely dangerous ones.

You will write prompts across benign, dual-use, and adversarial levels; evaluate model outputs against a policy standard; and craft the reference answer with the reasoning.

Qualifications

• Experience writing prompts and evaluating model responses in security-sensitive contexts.
• Ability to justify judgments with clear, accessible rationale for non-specialists.
• Strong written communication and analytical skills.

Responsibilities

• Write prompts across benign, dual-use, and adversarial levels.
• Evaluate model outputs against a defined policy standard.
• Draft the reference answer and technical reasoning for correct responses.

Skills

Prompts writing

Policy evaluation

Technical writing

Education

Experience in export control or proliferation analysis

Tools

Research methods

Job description

Mercor, based in New York, is assembling a panel of nuclear materials and safeguards experts to red-team frontier AI models. The aim is to test whether a model can judge misuse potential while answering legitimate questions fully and refusing genuinely dangerous ones.

You will write prompts across benign, dual-use, and adversarial levels; evaluate model outputs against a policy standard; and craft the reference answer with the reasoning.