About this role
OpenAI training work description
Create and review rigorous mathematics and statistics tasks for AI models, checking proofs, code, calculations, and reproducibility. Remote work pays $60 to $120 per hour equivalent, paid per task that meets the project specifications.
Generative AI & RLHF
Remote Per task · $60–$120/hr equivalent
$60–$120/hr
equivalent, paid per accepted task
Compensation
Contract, part-time
Engagement
Remote
Location
Nov 9, 2026
Posted
Open in 52 countries
United States + more
- Algeria
- Argentina
- Austria
- Bahrain
- Bangladesh
- Belgium
- Brazil
- Canada
- Chile
- Colombia
- Denmark
- Egypt
- France
- Germany
- Ghana
- Greece
- India
- Indonesia
- Italy
- Japan
- Jordan
- Kazakhstan
- Kenya
- Kuwait
- Kyrgyzstan
- Lebanon
- Malaysia
- Mexico
- Morocco
- Netherlands
- Nigeria
- Oman
- Pakistan
- Palestine
- Peru
- Philippines
- Portugal
- Qatar
- Saudi Arabia
- Singapore
- South Africa
- Spain
- Sri Lanka
- Switzerland
- Taiwan
- Thailand
- Tunisia
- United Arab Emirates
- United Kingdom
- United States
- Uzbekistan
- Vietnam
The work
You will help improve AI systems by creating difficult technical tasks and judging whether model answers are correct, well supported, and reproducible. The work combines mathematics, statistics, optimization, programming, and scientific computing.
- Create computational mathematics and statistics tasks for advanced AI models.
- Write clear instructions, prepare reference solutions, and define objective grading criteria.
- Build reproducible environments for technical tasks.
- Review proofs, calculations, programs, statistical results, models, and optimization outputs.
- Find logical errors, invalid assumptions, numerical instability, incorrect statistical methods, and reproducibility problems.
- Validate solutions against objective mathematical and technical standards.
What it pays and takes
This is a remote contractor role with a flexible schedule. The work is open to candidates in the listed eligible countries and requires professional fluency in English.
- Pay: $60 to $120 per hour equivalent, paid per task that meets the project specifications.
- Schedule: Less than 20 hours per week, with a flexible schedule.
- Work setting: Fully remote contractor role.
- Location: Open to candidates in eligible countries across the listed regions.
- Language: Professional fluency in English.
- Experience: Prior paid experience training, evaluating, benchmarking, or improving AI systems on mathematical or technical tasks.
- Technical skills: Strong practical proficiency with Python and hands-on experience with computational methods.
- Expertise: Deep knowledge in at least one area such as applied mathematics, scientific computing, operations research, optimization, probability, statistics, numerical methods, theorem proving, or mathematical modeling.
- Judgment: Ability to assess technical correctness, assumptions, and reproducibility.
Helpful background
Experience creating AI-training tasks, reference answers, rubrics, graders, or benchmark problems can support your application. Reviewing technical submissions from other AI trainers is also useful.
- Peer-reviewed publications or preprints.
- Public mathematical or statistical code.
- Open-source software contributions.
- Research profiles, competition results, or reproducible notebooks.
- Models, proofs, benchmarks, or comparable technical projects.
- Familiarity with NumPy, SciPy, SymPy, pandas, Pyomo, OR-Tools, Lean, Coq, or similar tools.
- An advanced degree or an equivalent record of substantial computational work.
How it works
You apply here in about two minutes and finish on the hiring platform. The next step is a short AI interview (technical roles include a coding challenge). Passing it is how you qualify for paid tasks, so take it soon after you are invited.
The hiring platform's application for this job asks for your resume, your LinkedIn profile, your phone number, your location and answers to 8 screening questions.
About AI training work
AI training is the human work behind systems that generate and analyze content. Contributors create examples and review model outputs, while people with strong technical expertise help set reliable standards for difficult problems.
Requirements
- Experience: Intermediate level
- Languages: English