You will create and evaluate difficult technical problems that test whether AI systems can reason through real engineering and physical science work. Problems should use realistic constraints, supporting data, and technical judgment rather than rely on a textbook answer.
You will also review other contributors’ work, check whether solutions are technically sound, and identify problems that are ambiguous or cannot be graded consistently.
- Design challenging problems in your engineering or physical science specialty.
- Provide a correct solution with clear reasoning that another expert can verify.
- Set objective criteria so reviewers can reach consistent correctness decisions.
- Review submitted problems for technical accuracy and sound methods.
- Flag problems that are unclear or cannot be graded reliably.
What it pays and takes
This is flexible, fully remote, part-time contractor work for experienced technical professionals. The role supports contributions alongside your current work.
- Pay: $85 to $185 per hour.
- Work arrangement: Fully remote, flexible, part-time contractor role.
- Location: Open to candidates in the countries listed for this role across the Americas, Europe, Asia, Africa, and the Middle East.
- Language: Professional fluency in English.
- Education: A master's or PhD in engineering, physics, materials, or a closely related field.
- Experience: Principal-, staff-, or fellow-level hands-on ownership of technical work at a major engineering organization or national lab, or a faculty appointment at a leading university.
- Expertise: Deep knowledge in at least one relevant engineering or physical science area.
- Skills: Strong technical judgment, clear explanations, and the ability to validate solutions.
- Helpful background: Publications, patents, or technical leadership showing practical ownership of complex engineering or scientific work.
How it works
Apply on OpenTrain with your resume, then complete the application on the hiring site.
About AI training work
AI training is the human work behind modern AI systems, including writing and reviewing problems, rating model answers, and checking whether outputs are correct. People with deep technical expertise help define the difficult examples and reliable standards that AI systems need to improve.