Get paid to evaluate AI answers with your medical knowledge
AI companies need physicians and other healthcare professionals to check whether their models answer medical questions correctly. The work is remote, freelance and on your own schedule. Here are the three platforms I work on myself, and how to get in.
The work in practice
- Check AI answers to medical questions: is the diagnosis plausible, the dose correct, the advice in line with guidelines?
- Grade answers against criteria and explain the errors.
- Write hard clinical questions that models get wrong.
- Build grading rubrics for clinical scenarios.
- Review medical translations and texts in your native language.
- No patients, no on-call shifts, no treatment liability.
The three platforms at a glance
| Platform | How it works | Onboarding | Good fit for | |
|---|---|---|---|---|
| Mercor | Specific project roles with a fixed hourly rate. You apply to individual roles. | CV and a short AI interview, sometimes a domain test. | Specialists, focused projects | Go to Mercor |
| micro1 | Open roles for experts, usually hourly or per task. | Profile and AI interview, then matching to projects. | Physicians of all specialties, multilingual roles | Go to micro1 |
| Outlier | Task-based work on one platform, paid hourly on projects. | Sign-up, skill assessments and short trainings per project. | Getting started, also medical students and nurses | Go to Outlier |
Rates and projects change constantly and depend on specialty, language and country. You see current offers after signing up. Tip: join more than one platform so you have work more often.
How to get in
Create a profile
Upload an English CV. State your specialty, licenses and languages clearly.
AI interview
A short video or audio conversation. Answer in a structured way, like a case presentation.
Domain test
Medical questions or grading sample answers. Read the instructions carefully.
First project
Once approved you get tasks. Quality matters more than speed.
Tips for the assessments
- Project guidelines beat clinical gut feeling. Grade exactly by the criteria.
- Justify every rating: which error, why it is dangerous, what would be correct.
- Cite guidelines where you can (e.g. ESC, NICE).
- Take your time. Rushed tests often get you locked out of a project.
- Don't use AI during the test if it is not allowed. Platforms check for it.
- List every language you speak. Medical tasks in languages other than English are especially in demand.
Who writes this page
I hold a medical license in Austria and am also licensed in Sweden and Norway. I live in Portugal and have been evaluating medical AI answers for Mercor, micro1 and Outlier since February 2026. Here I share what I've learned.
Mercor, micro1 or Outlier?
My honest take on which platform is most worthwhile for physicians, and what I have learned since February 2026.
FAQ
Do I need a current medical license?
Many expert roles require a medical degree, some a specialist qualification. There are also roles for students, nurses and pharmacists.
Is this employment?
No. You work as a freelancer and pay tax on the income in your country of residence. Ask a tax advisor if unsure.
How much time do I need?
You decide. Some work a few hours a month, others twenty hours a week. The amount of tasks varies by project.
Is it compatible with my clinical job?
Check your employment contract for rules on side work. Often you have to report it.
The links to Mercor, micro1 and Outlier are personal referral links. If you sign up through them and work there, the platform pays me a bonus. It costs you nothing and does not reduce your pay.
This site is independent. I am not an employee, recruiter or representative of any of the platforms. I cannot guarantee acceptance, projects or earnings.