Scientific Coding Task Trainer

You will create scientific programming tasks used to train and evaluate AI models. Each task should turn a challenging STEM idea into a clear coding problem with a reliable way to judge the answer.

You will also build verified Python solutions, write complete unit tests, and improve tasks based on quality feedback and model evaluation results.

• Write one main scientific problem with at least three connected sub-problems.
• Implement verified Python reference solutions.
• Create full unit test coverage and test cases that separate correct from incorrect outputs.
• Check tasks for clear requirements, scientific accuracy, consistent results, and strong test quality.
• Refine tasks after feedback from language model evaluations.
• Join review discussions, feedback sessions, and project standups during overlap hours.

What it pays and takes

This is a remote contractor assignment for candidates in a defined group of countries. Compensation is not disclosed.

The role requires advanced STEM knowledge and strong Python skills. Prior AI annotation, research, or scientific writing experience is useful but not required.

• Pay: Compensation is undisclosed.
• Schedule: 40 hours per week for an eight-week contract, including four hours of overlap with Pacific Time. The listing also records a minimum commitment of 20+ hours per week.
• Location: Bangladesh, Brazil, Colombia, Egypt, Ghana, India, Pakistan, Indonesia, Kenya, Nigeria, Turkey, or Vietnam.
• Language: English.
• Education: A master's or PhD in mathematics, physics, chemistry, biology, computer science, or a related STEM field.
• Required skills: Strong Python programming, scientific computing, rigorous problem writing, clear constraints and expected outputs, and careful evaluation of correctness, determinism, and test quality.
• Helpful background: LLM evaluation frameworks, coding benchmarks, NumPy, SciPy, SymPy, published research, or academic STEM projects.

How it works

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

About AI training work

AI training work is the human side of building AI: people create examples, review answers, and test whether models follow instructions correctly. Scientific coding specialists are needed because reliable evaluation depends on accurate reasoning, working code, and tests that catch subtle errors.