Clinical work is among the best paid on this site. The median open role advertises a ceiling of $170 an hour, against $100 across the site as a whole. Only QA engineering has a higher median, across seven roles.
There are 26 open roles carrying 1,268 openings, from Mercor (14), micro1 (7), Terac (4) and AfterQuery (1). Hourly rates span $50–400.
It is a small category. High rates and few seats is the pattern throughout this market, and medicine is a clear example of it.
What the top of the range looks like
Mercor lists a Radiology Expert role at $300 to $400, the highest hourly band in the category, and a US-licensed Dermatologist role at $270. Its Physician Talent Network runs $110 to $250.
micro1 lists a Clinician role in oncology and haematology at $160 to $200, and roles for military and veteran mental health experts and bilingual mental health experts, both at $100 to $200.
Mercor also runs a cluster of multilingual physician roles focused on clinical documentation: outpatient and primary care at $170 to $190, and inpatient hospitalist work at $170. Clinical fluency in a second language is part of what those roles pay for.
How much do paid medical interviews pay?
4 of the 26 clinical roles pay a fixed amount per session, not by the hour, and Terac carries most of them.
Open now are an evaluation of AI diagnostic outputs aimed at dermatologists at $210 for the session, a paid interview with practising clinicians on outpatient care at $210, an interview for nurse practitioners and physicians on payer-sponsored in-home care at $200, and an interview with US healthcare professionals on medication selection at $175.
A session fee is not an hourly rate. The listings state a fee and no duration, so this site does not convert them, and neither should you when comparing. What they represent is a defined, short commitment, not ongoing work.
What the work involves
The category divides into three kinds of task.
Evaluating model output against clinical judgement. Reading what a model produced for a case and deciding whether it is correct, safe, and reasoned the way a clinician would reason. The dermatology and radiology roles sit here.
Clinical documentation. The multilingual physician roles are explicit about this: producing or correcting documentation so a model learns how a clinician actually records an encounter.
Structured interviews and surveys. The Terac sessions, where the deliverable is your account of how decisions get made in practice.
Across all three, the thing being bought is the judgement a licence represents. A model can generate a plausible differential; what it cannot do is know which plausible answer is wrong in the way that matters.
Can you do this outside the US?
Of the 26 roles, 0 state a US-only restriction outright, though several name US licensure in the title or scope, and the payer-focused research sessions are inherently about the US system.
The multilingual roles run the other way, and treat non-English clinical fluency as the qualification.
What to settle before taking one
Patient information stays out. These contracts want your reasoning, not your records. Nothing in this work requires identifiable patient data, and any engagement that seems to invite it is worth stopping over rather than working around.
Your licence and your employer. Whether contract work of this kind interacts with your employment terms, your licensure conditions or your indemnity cover is specific to your jurisdiction and your position. This site cannot advise on any of it.
What your judgement becomes. Your reasoning is being reused to shape how a model answers clinical questions for people you will never see, without you in the loop. That is legitimate and well paid, and it is also a different thing from treating a patient. Whether that sits comfortably is a judgement worth making deliberately rather than discovering later.
Where to start
Browse the open clinical roles to see current listings with their stated rates and units, or read what AI training work pays for how medicine compares with the rest of the site.