Mathematics Expert (Master’s/Ph.D.)

You will help evaluate advanced mathematics language models using problems, model-generated solutions, computational tasks, and formal proofs. The work covers topics from early undergraduate through doctoral-level mathematics.

• Design original, challenging problems that test multi-step, abstract, and proof-based reasoning.
• Solve mathematics problems independently and write clear, logically structured solutions with full justifications.
• Review model-generated solutions, find mathematical errors or missing arguments, and provide precise feedback, annotations, and corrections.
• Build and validate Python solutions with approved scientific libraries.
• Translate mathematical problems and proofs into Lean and verify that the formal proofs compile correctly.

This is a fully remote contractor assignment. The role details do not list a pay rate.

• Time commitment: At least 20 hours per week, with options for 20, 30, or 40 hours per week.
• Daily schedule: At least 4 hours per day, including 4 hours of overlap with Pacific Time.
• Education and background: A master’s or Ph.D. in mathematics, or relevant academic or professional expertise in mathematics, applied mathematics, statistics, or a related field.
• Mathematics knowledge: Strong foundations in areas such as algebra, calculus, analysis, geometry, or topology.
• Communication: Professional fluency in English and the ability to explain complex mathematical ideas clearly with structured language and examples.
• Work style: Strong attention to detail, independent work habits, constructive feedback skills, and the ability to follow detailed project guidelines.
• Useful technical skills: Python and scientific libraries are valuable; familiarity with Lean or formal proof verification is beneficial.

Apply on OpenTrain with your resume, then complete the application on the hiring site.

AI training is the human work behind systems that learn from examples, including written answers, ratings, code, and formal solutions. Mathematics experts help make these examples accurate and useful so models can reason more reliably.