This is a fully remote position, open to applicants in United States.

📋 Description

• Craft challenging single-turn prompts related to nuclear materials and safeguards, categorized as benign, dual-use, or adversarial.

• Assess model outputs based on an established policy standard.

• Evaluate whether the responses accurately addressed technical inquiries.

• Develop reference answers and clarify the technical reasoning behind them.

• Conduct red-team exercises on frontier AI models regarding potential misuse judgments.

• Produce written rationales that can be easily understood by non-specialists.

• Engage in reading and writing about misuse scenarios for extended periods.

⛳️ Requirements

• Experience in characterizing actual nuclear materials.

• Background in nuclear forensics casework or international exercises, including radiochemical and mass-spectrometric analysis, detection and interdiction technology, material signature libraries and attribution methodology, or pre-/post-detonation forensics research.

• Capability to differentiate between routine professional inquiries and potentially hazardous requests.

• Exceptional technical writing skills.

• Proficiency in writing rationales that are comprehensible to non-specialists.

• Previous experience in technical writing, published research, or serving as an expert witness is highly regarded.

• Must refrain from utilizing classified or export-controlled information, NDA-covered information, or information subject to prepublication review requirements.

• Applicants with such obligations should disclose them to ensure appropriate scoping of the work.