Mathematics Expert, Advanced Problem Solving & Evaluation
About OpenTrain
OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We help contributors discover projects, consolidate their work history, and grow a durable freelance career teaching AI systems how to reason.
Why AI training matters for mathematicians
AI models learn from human examples and explanations — and mathematical reasoning is one of the most important, high-impact areas where expert human judgment shapes model behavior. Contributors turn advanced math knowledge into datasets and evaluations that directly improve how models solve problems, prove statements, and explain ideas.
This work is highly flexible, remote, and accessible: many projects let you set your own hours while contributing at the cutting edge of how AI systems handle rigorous reasoning.
The role
As a Mathematics Expert for OpenTrain you'll produce and review advanced mathematical content used to train and benchmark generative models. You will solve problems, write detailed proofs and technical analyses, evaluate AI-generated reasoning for correctness and rigor, and create annotated datasets with clear documentation.
This is a contract, part-time role requiring 20+ hours per week. Work is fully remote and open worldwide; all work and deliverables are in English. Compensation is hourly at $20–$40/hr depending on task complexity and your experience.
- Employment type: Contractor, part-time
- Time requirement: 20+ hours per week
- Pay: USD $20–$40 per hour
- Work location: Remote, worldwide
- Primary language: English
What you'll do day to day
Your core tasks will center on producing high-quality mathematical content and evaluating model outputs so they meet standards of correctness and clarity.
- Review and solve advanced problems in pure and applied mathematics
- Write detailed solutions, proofs, and technical analyses suitable for training data
- Evaluate AI-generated mathematical reasoning for accuracy, logical soundness, and clarity
- Annotate and curate domain-specific text datasets for model training and benchmarking
- Document methodologies, edge cases, and evaluation criteria in clear written form
- Collaborate remotely with other subject-matter experts to align standards
Requirements
You must bring a strong mathematics background and demonstrated experience communicating technical mathematics in writing.
- BSc, MSc, or PhD in Mathematics, Applied Mathematics, or a closely related field
- Experience authoring research papers, mathematical proofs, or technical reports
- Proven ability to write clear, rigorous proofs and explanations
- Experience evaluating mathematical reasoning, logic, or computational solutions
- Excellent written and verbal communication in English, with precision and clarity
- Familiarity with current trends, open problems, or applications in mathematics
Who should apply
This role suits mathematicians and quantitative researchers who enjoy teaching and critiquing reasoning in written form. It is open to early-career and experienced candidates who meet the educational and communication requirements.
- Graduate students, postdocs, instructors, researchers, or industry mathematicians
- People who take care to write rigorous, well-explained proofs and analyses
- Contributors who prefer flexible, remote, project-based contract work
How the work is structured
Tasks are delivered as text-based annotations, solutions, and evaluation ratings. Labeling work includes classification, text generation, evaluation/rating, and data-collection style contributions focused on mathematical content. You will be paid hourly and submit work through OpenTrain's workflow for review and quality control.
OpenTrain handles contracting and payment; you maintain control of your profile and portfolio as you complete projects.
- Primary data type: Text
- Label/annotation types: Classification, Text generation, Evaluation rating, Data collection
- Payment type: Hourly (PAY_PER_HOUR), $20–$40/hr
- All assignments and quality requirements are provided with every task