Analytical Chemist – AI Safety Red-Teaming Specialist

About the role

We're building a panel of expert analytical and safety-focused chemists to help evaluate and stress-test frontier AI models. Your job is to probe whether an AI system can correctly distinguish between legitimate technical requests and requests that carry genuine misuse potential — and to define, in writing, what a correct and responsible answer looks like.

This is a remote, flexible-hours engagement ideal for chemists who enjoy applying deep domain expertise to a novel, high-impact problem space: making advanced AI systems safer and more reliable when handling chemistry-related queries.

What you'll do

• Draft challenging, single-turn prompts within your area of chemistry expertise, spanning three risk tiers: benign, dual-use, and adversarial.
• Review AI-generated responses to these prompts against a defined safety and accuracy policy.
• Judge whether each response was handled appropriately — correctly answering legitimate questions in full while appropriately declining or redirecting genuinely dangerous ones.
• Author clear, well-reasoned reference answers that model what a correct, policy-compliant response should look like.
• Provide structured written feedback and rationale to support model evaluation and improvement efforts.
• Collaborate with a distributed panel of subject-matter experts and program coordinators to refine evaluation standards over time.

Requirements

• Strong background in analytical chemistry, chemical safety, or a closely related discipline (industry, academic, or research experience).
• Demonstrated understanding of dual-use chemical knowledge and the ability to reason carefully about misuse potential.
• Excellent written communication skills — able to produce clear, precise, and well-structured technical prompts and reference answers.
• Sound judgment and attention to detail when evaluating nuanced or ambiguous technical content.
• Comfort working independently in a remote, asynchronous environment.
• Prior experience with content review, red-teaming, or AI evaluation work is a plus but not required.

Compensation

This role is compensated hourly, in the range of $65–$75/hour, commensurate with experience and scope of contribution. Engagement is remote and flexible, with opportunities for ongoing work based on performance and project needs.