Mathematics AI Response Evaluation Specialist

About OpenTrain

OpenTrain is the centralized, open platform where people build careers in AI training and data labeling. We help freelancers find projects, collect credible proof-of-work, and grow a long-term portfolio of AI-training experience.

OpenTrain AI is the hiring and contracting organization for this role. Contributors work remotely and build reputation by doing real annotation and evaluation work that directly shapes how AI systems behave.

About AI training work

AI training (also called data labeling or RLHF) is the human side of building AI: people create, check, and rate examples that models learn from. This role focuses on mathematical content — evaluating solutions, reasoning, and correctness.

These projects are often flexible and remote, making them suitable for part-time work alongside research or other jobs. Contributors influence model behavior and help make AI systems more reliable on quantitative tasks.

The role — what you'll do

You will read AI-generated mathematical responses and evaluate them for correctness, clarity, and reasoning quality. Your assessments will include checking calculations, proof structure, methodology, and whether conclusions follow from premises.

You will also write or refine clear model solutions and explanations, and compare multiple AI answers to identify which is strongest mathematically and logically.

• Review AI text responses for correctness, reasoning quality, and clarity.
• Detect calculation errors, unjustified steps, and domain-switching mistakes.
• Validate proofs, modeling choices, probability/statistics arguments, and optimization reasoning.
• Write high-quality mathematical explanations and model solutions.
• Rank and compare responses using evaluation ratings and written justification.

Requirements

This role requires advanced formal training and strong mathematical communication. Preserve attention to the listed qualifications exactly when you apply.

• MS or PhD in mathematics, statistics, or a related field from a top 100 university.
• Strong command of pure and applied mathematics, proofs, modeling, probability, statistics, and optimization.
• Advanced proofreading judgment for mathematical arguments and the ability to detect calculation errors and unjustified steps.
• Ability to fact-check quantitative claims and validate mathematical reasoning.
• Excellent English writing skills and clear mathematical explanation.
• Background in research, analytical writing, debate, programming, or mathematics.
• Prior AI model evaluation, RLHF, or data-labeling experience is a plus.

Helpful background

You will stand out if you have experience developing or critically reviewing complex math content such as problem banks, textbook sections, proofs, or research notes. Prior work producing model solutions, grading or tutoring advanced math, or peer-reviewing technical material is highly relevant.

Logistics, pay, and how to apply

This is a contract, part-time role (20+ hours/week) paid hourly at USD 70. Work is remote but limited to the following countries listed in the posting.

The task is text-based evaluation and labeling: you will read AI-generated text, assign evaluation ratings, and provide written justification and model solutions. Apply through your OpenTrain profile to be considered; building a strong OpenTrain portfolio helps you qualify for more projects over time.

• Compensation: USD 70 per hour, paid per hour as a contractor.
• Time: 20+ hours per week (part-time contractor).
• Languages: English required.
• Allowed work locations: Vietnam (VN), Timor-Leste (TL), Thailand (TH), Singapore (SG), Pakistan (PK), Philippines (PH), Nepal (NP), Malaysia (MY), Sri Lanka (LK), Cambodia (KH), Indonesia (ID), Bhutan (BT), Bangladesh (BD), India (IN), Brazil (BR), Germany (DE), United States (US).
• Data type: Text — tasks are evaluation/rating of AI responses (EVALUATION_RATING).