About the work
Mercor is assembling a panel of explosives and energetic materials experts to red-team frontier AI models. The goal is to test whether a model can correctly judge the misuse potential of a technical request: Answering legitimate questions fully while refusing genuinely dangerous ones.
You will
- Write challenging single-turn prompts in your domain, labelled across three levels: benign, dual-use, and adversarial.
- Evaluate the responses that come back against a defined policy standard, and judge whether each was handled correctly.
- Write the reference answer; what a correct response looks like, and the technical reasoning for why.
Models tend to fail this domain in both directions: refusing routine questions a mining engineer or investigator asks daily, then engaging with a request they should have shut down because it arrived in academic clothing. Knowing which is which takes someone who has done the work.
Background we're looking for
We're recruiting across two profiles and need both: chemists with specialty materials expertise, and engineers or operators with initiation and propulsion system experience. Prior AI evaluation or red-teaming experience is preferred. Relevant backgrounds include:
- Hazardous Device Specialist / Public Safety Professional — render-safe procedures and device assessment.
- Specialty Materials Chemist (pyrotechnics or propulsion) — formulation, synthesis and performance testing.
- Federal Certified Materials Specialist / Incident Analyst — forensic identification and scene analysis.
- Propulsion & Effects Engineer — munitions and propulsion system design.
- Commercial Blasting Engineer — specialty materials handling, storage and transport.