Math AI Trainer

MissionHires; United States; Remote; USD 55,000 - 172,000; Full time

Benefits offered by this job

Hourly contract

Fully remote

Flexible scheduling

Minimum 10 hours/week

Opportunities for 30-40+ hours/week

Paid project milestones

AI tools access

Regular paid projects

Job summary

MissionHires is seeking a detail-oriented contractor to design and evaluate AI tasks for professional STEM workflows. You will create realistic datasets, scenarios, and grading rubrics, and work with frontier AI models to assess accuracy against professional standards.

This fully remote, hourly-contract role offers flexible scheduling, a minimum of 10 hours per week, and opportunities for more hours as projects scale.

Qualifications

• In-progress Bachelor's degree or higher in mathematics, physics, chemistry, biology, an engineering discipline, data science, or a related field.
• Two or more years of applied professional experience preferred, ideally outside undergraduate study.
• Current or recent hands-on individual-contributor experience in a scientific, engineering, or data science discipline.
• Ability to work with real professional data and domain-specific materials, including instrument or test data, measurement files, simulation outputs, schematics, drawings, reports, and related analysis tools.
• Strong technical judgment and ability to identify incorrect assumptions, unit errors, scaling errors, invalid boundary conditions, unsupported conclusions, and flawed interpretations.
• Comfort with statistical and experimental reasoning, including experiment design, measurement error, uncertainty, and common statistical pitfalls.
• Strong written and spoken English, with the ability to explain why a result is incorrect rather than only identifying that it is incorrect.
• General familiarity with AI and LLM tools and the ability to distinguish sound reasoning from plausible but inaccurate output.
• Working understanding of adjacent technical sub-disciplines sufficient to assess work outside a primary specialty.
• Coding or scientific computing skills in Python or a similar language are strongly preferred and required for certain projects.
• Comfort working at the command line and authoring or verifying technical work in code for applicable projects.
• Baseline technical literacy with cloud file tools, browser profiles, desktop application installation, file conversion, and compressed-file handling.
• Advanced degree or professional licensure is a plus but not required.

Responsibilities

• Design realistic, challenging technical tasks based on professional STEM, engineering, or data science workflows, including scenarios, colleague-style prompts, and self-authored supporting files.
• Create and assemble professional materials such as test data, measurement files, specifications, drawings, schematics, simulation outputs, lab records, reports, and analysis files.
• Run tasks through frontier AI models and evaluate calculations, analyses, design reviews, reports, and other deliverables against professional standards.
• Compare multiple model outputs for the same prompt and source materials, determine which output performs better, and document shortcomings in each response.
• Write detailed grading rubrics defining required assumptions, methods, units, magnitudes, acceptance criteria, failure modes, and other elements of a correct response.
• Document specific model failures, including unit or scaling mistakes, data misinterpretation, unsupported conclusions, fabricated or ignored source files, missed safety or boundary conditions, and off-brief responses.
• Review, refine, and provide feedback on tasks developed by other subject-matter experts.
• Propose task scenarios from professional experience, identify likely model failure points, and build accepted proposals into complete tasks.
• Work independently in a structured, long-form remote environment while producing clear written reasoning and evidence-based evaluations.

Skills

Strong written English

Strong verbal English

Technical judgment

Statistical reasoning

Python programming

Command line proficiency

Cloud file tools

Independent work

Education

Bachelor's degree or higher in math/physics/chemistry/biology/engineering/data science or related field

Tools

Python

Job description

When AI gets the details wrong in your field, real people pay for it. The people who know the work best have to be the ones training these models. You're not being replaced by AI. You're making sure it's used accurately and safely.

Responsibilities

• Design realistic, challenging technical tasks based on professional STEM, engineering, or data science workflows, including scenarios, colleague-style prompts, and self-authored supporting files.
• Create and assemble professional materials such as test data, measurement files, specifications, drawings, schematics, simulation outputs, lab records, reports, and analysis files.
• Run tasks through frontier AI models and evaluate calculations, analyses, design reviews, reports, and other deliverables against professional standards.
• Compare multiple model outputs for the same prompt and source materials, determine which output performs better, and document shortcomings in each response.
• Write detailed grading rubrics defining required assumptions, methods, units, magnitudes, acceptance criteria, failure modes, and other elements of a correct response.
• Document specific model failures, including unit or scaling mistakes, data misinterpretation, unsupported conclusions, fabricated or ignored source files, missed safety or boundary conditions, and off-brief responses.
• Review, refine, and provide feedback on tasks developed by other subject-matter experts.
• Propose task scenarios from professional experience, identify likely model failure points, and build accepted proposals into complete tasks.
• Work independently in a structured, long-form remote environment while producing clear written reasoning and evidence-based evaluations.

Requirements

• In-progress Bachelor's degree or higher in mathematics, physics, chemistry, biology, an engineering discipline, data science, or a related field.
• Two or more years of applied professional experience preferred, ideally outside undergraduate study.
• Current or recent hands-on individual-contributor experience in a scientific, engineering, or data science discipline.
• Ability to work with real professional data and domain-specific materials, including instrument or test data, measurement files, simulation outputs, schematics, drawings, reports, and related analysis tools.
• Strong technical judgment and ability to identify incorrect assumptions, unit errors, scaling errors, invalid boundary conditions, unsupported conclusions, and flawed interpretations.
• Comfort with statistical and experimental reasoning, including experiment design, measurement error, uncertainty, and common statistical pitfalls.
• Strong written and spoken English, with the ability to explain why a result is incorrect rather than only identifying that it is incorrect.
• General familiarity with AI and LLM tools and the ability to distinguish sound reasoning from plausible but inaccurate output.
• Working understanding of adjacent technical sub-disciplines sufficient to assess work outside a primary specialty.
• Coding or scientific computing skills in Python or a similar language are strongly preferred and required for certain projects.
• Comfort working at the command line and authoring or verifying technical work in code for applicable projects.
• Baseline technical literacy with cloud file tools, browser profiles, desktop application installation, file conversion, and compressed-file handling.
• Advanced degree or professional licensure is a plus but not required.

Benefits

• Hourly contract compensation of $40 to $125 USD per hour, paid through PayPal on a regular cadence.
• Fully remote work with flexible scheduling.
• Minimum commitment of 10 hours per week, with no weekly maximum and opportunities for 30 to 40 hours per week or full-time hours.
• Multi-day task timers that allow work to be spread across several days.
• Access to Claude and ChatGPT through the project; no personal subscription is required.
• Regular paid projects in the contributor's technical field after successful assessment and onboarding.
• Compensation for completed stages of the task-development process, subject to acceptance requirements.
• Opportunity to help improve frontier AI systems using professional science, engineering, and data science expertise.