Senior Physicist, AI Reasoning & Benchmark Design (Remote)

Zekai; United States; À distance; USD 180 000 - 260 000; Plein temps

Résumé du poste

Zekai is seeking a senior physicist to help frontier AI models reason about real physics. You will review physics reasoning tasks and model outputs, write clear problem statements and reference solutions, and design benchmarks to measure genuine improvements.

This is a fully remote role with US-based applicants preferred. You should have a PhD in physics or related fields, a strong research track record, and hands-on experience with AI tools.

Qualifications

• PhD in physics, astrophysics, applied physics, or closely related field.
• Senior research role at a university, national lab, or research organization (faculty, PI, senior scientist).
• Genuine domain depth in at least one area (condensed matter, high-energy/particle, astrophysics, atomic/molecular/optical/quantum, or statistical/computational physics).
• Strong peer-reviewed publication record; fellowships/awards are a plus.
• Hands-on use of large language models in your work; ability to distinguish sound arguments from plausible but wrong ones.
• Excellent written communication and precise, structured feedback.

Responsabilités

• Expert review: Assess physics tasks and model outputs for rigor, avoiding flawed derivations and unjustified approximations.
• Problem specs and reference solutions: Write precise problem statements and rigorous reference solutions reflecting real research practice.
• Benchmarks: Design challenging, research-level problems and evaluation sets in your specialization area.
• Calibration: Collaborate with researchers in related fields to turn tacit expertise into explicit, teachable criteria.

Connaissances

AI fluency

Communication

Formation

PhD in physics/astrophysics/applied physics

Description du poste

Zekai is seeking a senior physicist to help frontier AI models reason about real physics. You will review physics reasoning tasks and model outputs, write clear problem statements and reference solutions, and design benchmarks to measure genuine improvements.

This is a fully remote role with US-based applicants preferred. You should have a PhD in physics or related fields, a strong research track record, and hands-on experience with AI tools.