$61-77/hrAs published by the platform.
This role involves writing and checking multiple-choice mathematics questions that serve as benchmarks for evaluating AI systems. It suits mathematicians with advanced training who can design rigorous problems or verify existing ones across several applied domains. The work is asynchronous and fully remote.
$61-77/hr
Pay as published by the platform. It is not a guarantee of income or hours.
Open to residents of: United States.
You apply on work.mercor.com: create a profile, complete a skills assessment, then get matched to projects that fit your background.
Official posting: https://work.mercor.com/jobs/list_AAABn_yIZR1__lQF4UFJWK4b/applied-mathematics-benchmark-specialist
Last checked on October 23, 2026.
This role involves translating mathematical proofs into machine-verifiable Lean 4 formalizations for an AI research organization. Mathematicians with formal proof system experience will work on bridging informal mathematical arguments and automated reasoning, identifying gaps in current verification tools and helping advance mechanized mathematics at the intersection of AI and rigorous proof.
This role involves evaluating and improving how AI models handle rigorous mathematical reasoning. You will create challenging math problems, review AI solutions for logical errors and unjustified steps, and write correct solutions that serve as training examples for the models to learn from.
This contract role asks a senior researcher in computer science or computational physics to guide frontier AI models and agents through multi-step research, from proofs and algorithms to numerical experiments and simulations. The person checks the results, judges their correctness and novelty, and writes memos on where the models succeed or fail. No prior AI experience is needed, since the domain expertise is the core requirement.
Posted October 13, 2026
Referral link to micro1's job list: search for this role there. Open this exact job
This role invites mathematical physicists to contribute to a research benchmark in fundamental physics, designed to assess whether cutting-edge AI models can solve real research problems in physics. You will work on specialized challenges in one of four proposed domains (special functions, integrable systems, permutation combinatorics, conformal geometry), creating, solving, reviewing, or validating problems at the level of academic research.
Posted September 25, 2026
This role involves contributing to the creation and evaluation of research-level physics challenges for a benchmark designed to test the physical reasoning capabilities of AI models. The work involves creating problems, solving them, examining solutions or auditing them in your area of specialization in statistical physics. This role is intended for PhDs in statistical physics with published experience on specific phenomena and who master the corresponding mathematical and computational methods.
Posted September 25, 2026
This ongoing project hires advanced STEM experts to write graduate-level and PhD-level problems with exact, verifiable answers, helping AI systems improve at rigorous scientific reasoning. Participants create original problems in their own specialty, write detailed step-by-step solutions, and sometimes review AI-generated answers to find reasoning errors. It is aimed at PhD candidates and holders, postdoctoral researchers, and faculty in biology, chemistry, physics, mathematics, and selected engineering fields.
Posted August 6, 2026
