Mathematics Research Collaborator — Proof Review & Evaluation

About OpenTrain AI

OpenTrain is the centralized platform where people build careers in AI training and data labeling. We connect expert contributors with high-impact projects that teach and evaluate state-of-the-art AI systems and help you grow a portable professional portfolio.

OpenTrain AI is hiring and contracting for this role. Our work focuses on the human side of AI development — the judgments, annotations, and expertise that shape model behavior — and we support flexible, remote engagements for qualified specialists.

Why AI training in mathematics matters

Modern AI models learn from examples and human feedback. For scientific and mathematical reasoning tasks, experts are needed to evaluate proofs, spot subtle errors, and create challenging problems that push models beyond textbook reasoning.

This role places you at the frontier of model evaluation and training: your assessments and problem sets will be used to measure and improve how systems construct, verify, and formalize mathematical arguments.

The role

As a Mathematics Research Collaborator you will review research papers, author and assess advanced problems, and evaluate model outputs on proof construction, verification, problem solving, formalization, and conjecture exploration.

This is senior-level, research-depth work for experts who can judge correctness and novelty in advanced mathematics and theoretical work — not textbook exercises.

• Employment type: Contractor, Part-time
• Time commitment: 20+ hours per week (remote)
• Location: United States only
• Language: English

What you will do

Work involves careful, expert judgment about subtle reasoning mistakes and research-level rigor. Tasks will include evaluating both human-authored and model-generated mathematics content and collaborating with other senior contributors on complex cases.

• Review mathematics research papers for rigor, novelty, and correctness
• Create and assess difficult mathematics problems for model training and evaluation
• Judge model outputs on proofs, verification, problem solving, formalization, and conjecture exploration
• Identify subtle reasoning errors such as unjustified steps, misapplied hypotheses, and hand-waved lemmas
• Collaborate with other senior contributors on frontier model evaluation work

Required qualifications

Candidates must demonstrate a strong, recent record of research and the ability to evaluate advanced mathematics at publication quality. You should be comfortable working independently and providing precise, documented judgments about correctness and novelty.

• Doctoral or postdoctoral training in mathematics or a related STEM field
• First-author or sole-author publications in top mathematics or adjacent theory venues
• Active research record in pure/applied mathematics, theoretical computer science, or statistics
• Experience writing or reviewing advanced mathematics problems
• Proven ability to spot subtle errors in proofs and reasoning
• Competitive fellowships, honors, grants, or olympiad-level distinction are strong signals of fit

Tasks, data types, and labeling work

You will work primarily with text data: research papers, proof sketches, problem statements, and model-generated arguments. Labeling tasks include evaluation ratings, question-answering style assessments, and text-generation review.

Label types you can expect: evaluation rating, question answering, and text generation assessments. Specific labeling software is not prescribed — training and task instructions will be provided.

• Data type: Text (research papers, proofs, problems, model outputs)
• Label types: EVALUATION_RATING, QUESTION_ANSWERING, TEXT_GENERATION

Compensation, schedule, and how to apply

Pay is hourly at contractor rates between $80 and $110 per hour; the listed top rate is $110/hr. The role supports remote, part-time work with an expectation of 20+ hours per week.

To apply, build an OpenTrain profile, include your CV and publication record, and highlight relevant research, honors, or olympiad distinctions. Selected contributors will receive task-specific training and guidelines before starting.

• Hourly pay: $80–$110 USD per hour (top rate $110/hr)
• Schedule: Flexible, part-time contractor engagement (20+ hrs/week)
• Location requirement: United States