Mercor seeks Physics PhDs specialising in Advanced Quantum Mechanics, Advanced Electrodynamics, or Advanced Classical Mechanics for a premier project with one of the world’s top AI labs. You will contribute deep subject matter expertise to a cutting-edge AI training project involving state-of-the-art large language models, specifically by generating difficult, domain-specific problems that current frontier models struggle to solve correctly. The work is problem-writing and quality evaluation rather than annotation — you are the expert, not the reviewer. Tasks run at 4–6 per week, each taking several hours, requiring rigorous physics expertise and the ability to follow complex instructions precisely. Your contributions directly inform the training data shaping the next generation of AI systems. This is a fully remote, ongoing role starting in February with rolling applications. Experts are expected to contribute consistently over the project term.

Mercor seeks Physics PhDs specialising in Statistical Mechanics, Condensed Matter Physics, or Atomic, Molecular and Optical (AMO) Physics for a premier AI training project with one of the world’s leading AI labs. You will contribute deep domain expertise by generating difficult, high-quality problems in your specialism — the kind that expose genuine reasoning failures in frontier AI models. This is not annotation work; it requires the ability to construct challenging, original problems and evaluate solutions with the rigour expected at a peer-review level. Tasks run at 4–6 per week, each requiring several hours of focused expert effort. The work requires rigorous physics expertise and an ability to follow complex instructions precisely, including formatting requirements. Your contributions help create training data that will directly shape the quality of next-generation AI reasoning in your field. The role is ongoing from February with rolling applications and a fully remote structure.

This role focuses on evaluating and improving how AI models explain, reason about, and respond to physics questions across both foundational and advanced topics. Unlike the problem-writing roles above, the emphasis here is on assessing the quality of AI outputs rather than generating adversarial problems from scratch. You will write and refine prompts designed to guide model behaviour in physics contexts, evaluate LLM-generated responses for conceptual accuracy, mathematical correctness, and reasoning quality, conduct fact-checking using authoritative sources, and annotate model outputs by identifying strengths, inaccuracies, and areas for improvement. You will also assess clarity, structure, and appropriateness of explanations for different audience levels, and apply consistent evaluation standards using detailed rubrics and taxonomies. This role is broader in scope than pure problem-writing and is particularly well suited to physicists who are strong written communicators, have significant hands-on LLM experience, and are comfortable working to structured evaluation frameworks. Location is restricted to US, UK, and Canada — UK-based applicants are eligible.

This is not a specific project posting — it is an open application to join Mercor’s Physicist Expert Network. By applying here, you enter the pool from which Mercor matches physicists to AI training projects as they become available. The advantage of the talent network is flexibility: you apply once, complete your profile and pass an AI interview, and Mercor then invites you to interview for specific opportunities that match your expertise and schedule preferences, rather than requiring you to apply role by role. Projects in the network span a wide range of physics sub-disciplines and task types — from problem-writing and solution evaluation to prompt engineering and model annotation. Typical project commitments run from 15 to 30 hours per week. Payments are processed weekly via Stripe or Wise. If you hold physics expertise but are uncertain which specific role above best fits your background, or if none of the listed roles precisely matches your specialism, the talent network is the most practical entry point. It is also a good option if your schedule availability is variable and you want to be considered for shorter or part-time project engagements as they arise.

Mercor seeks Physics PhDs specialising in General Relativity, Astrophysics, or Cosmology for a premier project with one of the world’s top AI labs. You will contribute deep subject matter expertise to a cutting-edge AI training project involving state-of-the-art large language models, specifically by generating difficult, domain-specific problems that current frontier models struggle to solve correctly. The work is problem-writing and quality evaluation rather than annotation — you are the expert, not the reviewer. Tasks run at 4–6 per week, each taking several hours, requiring rigorous physics expertise and the ability to follow complex instructions precisely. Your contributions directly inform the training data shaping the next generation of AI systems. This is a fully remote, ongoing role starting in February with rolling applications.