The work
You will help evaluate how AI models respond to safety-critical materials science questions. Your work will focus on finding the boundary between safe and unsafe requests and explaining each decision clearly.
- Write prompts and scenarios that test whether materials science requests are safe or unsafe.
- Review and classify prompts created by other experts.
- Explain why a prompt is acceptable or where it crosses a safety boundary.
- Revise your analyses in response to structured reviewer feedback.
What it pays and takes
- Pay: $85 per approved task.
- Work type: Remote, flexible, task-based contractor work.
- Location: Open to candidates in Australia, Canada, the United Kingdom, New Zealand, or the United States.
- Education: Master's degree or higher in materials science, materials engineering, metallurgy, or a closely related field.
- Communication: Explain the technical reasoning behind safety judgments clearly and precisely.
- Helpful background: Experience with AI evaluation, red teaming, or data annotation.
- Valued expertise: A PhD or professional experience in metallurgy, advanced or energetic materials, nuclear materials, or national laboratory research.
- Language: Fluent English.
How it works
Apply on OpenTrain with your resume, then complete the application on the hiring site.