You will turn mathematical and research workflows into self-contained, terminal-based tasks that test AI agents. Your work will cover numerical analysis, optimization, statistics, mathematical modeling, probability, dynamical systems, numerical integration, differential equations, matrix computation, stochastic modeling, and algorithm analysis.
- Build tasks with datasets, equations, model definitions, constraints, and expected outputs.
- Implement expert solutions in Python, R, Julia, C/C++, Bash, or another relevant language.
- Create grading criteria covering accuracy, convergence, complexity, feasibility, and mathematical correctness.
- Define tolerances, stopping criteria, stability requirements, and reproducibility controls.
- Write automated tests for edge cases and alternative valid implementations.
- Debug floating-point precision, solver, conditioning, convergence, and performance issues.
- Document assumptions, mathematical formulations, expected outputs, and known limitations.
What it pays and takes
This is a remote contractor assignment for five weeks. The role requires advanced technical experience in mathematics, statistics, or a closely related field, along with the ability to independently implement, test, and validate computational algorithms.
- Pay: $300 per approved task.
- Schedule: At least 6 hours per day and 40 hours per week.
- Time overlap: At least 4 hours overlapping with Pacific Standard Time.
- Location: Remote and open worldwide.
- Language: English.
- Background: Ph.D., postdoctoral experience, or equivalent advanced technical experience in mathematics, statistics, or a closely related discipline.
- Programming: Strong ability in Python, R, Julia, C/C++, Bash, or another relevant language.
- Environment: Experience working in Linux or terminal-based environments.
- Expertise: Practical experience in numerical methods, optimization, statistics, mathematical modeling, or scientific computation.
How it works
Apply on OpenTrain with your resume and then complete the application on the hiring site.
About AI training work
AI training is the human work behind systems that learn from examples, including writing, testing, and evaluating model outputs. This role uses specialized mathematical and programming knowledge to create reliable evaluations that show whether AI agents can solve computational problems correctly.
