Most people who don't get in fail the screening, not the CV. Five minutes here first: How to pass the micro1 AI interview with Zara.
What the work is
Opportunity Snapshot
Pay: $45 - $90/hour
Commitment: Flexible — fits alongside your current role
You’ll need: A completed PhD in a STEM field, plus hands-on research experience
Your tasks: Build difficult technical problems from your own research and define what makes an answer correct
These tasks are hard to build, and that’s the point. We’re looking for people with genuine depth in their field: the problems where the textbook answer isn’t enough, where the right approach depends on reading the constraints correctly, where a plausible-looking answer is quietly wrong.
What you’ll do
- Design difficult technical problems drawn from your own research area, with realistic constraints and supporting data.
- Provide the correct solution and walk through the reasoning that gets there.
- Define objective criteria for correctness, so another researcher reviewing the work would reach the same verdict.
- Review problems built by other contributors for technical accuracy and soundness of method.
- Flag problems that turn out to be ambiguous or ungradable rather than submitting them.
Who gets hired
- Completed PhD in a STEM field. This is a hard requirement.
- Hands-on research experience — work you did yourself, not work you read about.
- Wet lab, research lab, or academic experience is strongly preferred.
- Able to explain why an answer is correct clearly enough that another researcher could verify it.
- Publications or a record of research contribution is a strong plus.
- Prior experience on AI training or evaluation projects is a plus, not a requirement.
How the selection process works
- Apply and complete the screening questions
- Complete the AI interview for this role — it’s how your expertise gets in front of the client
- Profile review
- Onboarding — our Talent Ops team will get you set up
The platform posted this on 7 Nov 2026.