Mercor is recruiting board-certified dermatologists for non-clinical AI evaluation work: interpreting clinical images with written reasoning, annotating lesions against a structured schema, and judging AI-generated dermatological assessments. Work is remote and asynchronous, drawn from a shared pool of concurrent workstreams, at a minimum of 15 hours per week with an observed rate around $270/hr.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work actually involves

You are not diagnosing patients — you are defining what a defensible dermatological assessment looks like so it can be measured. On a given day that might mean working through a case that pairs images with history and writing out your reasoning chain rather than just the impression; describing a lesion against a fixed schema (morphology, colour, configuration, distribution, body site) where the vocabulary is constrained and consistency matters more than eloquence; recording a ranked differential with justification for each item; or reviewing a model's written assessment and marking where it invented a finding, missed one, or stated a borderline call with unearned confidence.

Senior contributors also write the rubrics — the criteria that separate an excellent assessment from a merely acceptable one — and author difficult cases designed to probe where model reasoning breaks down, including presentations that look materially different across skin tones. Tasks run from roughly 10 minutes for structured annotation to about an hour for full case authoring, and each stream comes with its own throughput target.

What the screen is looking for

Mercor's screening is AI-led and verification-heavy. Expect it to confirm residency completion, board certification (ABD or a jurisdictional equivalent, or board-eligible with residency complete), an active unrestricted licence, and 3+ years post-residency practice. Beyond credentials, it probes descriptive precision — whether you can characterise a lesion in terminology another dermatologist would reproduce — and evaluation judgment: whether you can articulate why an assessment is wrong, not merely that it is. Candidates who default to "I'd biopsy" or "I'd need to see it in person" without then reasoning from the image available tend to stall, because the work requires committing to a graded judgment under image-only constraints.

Logistics

Fully remote, asynchronous, self-scheduled within stream deadlines
Minimum 15 hours per week, with capacity to concentrate hours when a stream is time-boxed
Shared expert pool: you may be matched across workstreams by subspecialty, availability, and interest, and may move between them
Non-clinical throughout — no patient care, no live diagnostic responsibility
$270/hr is the observed band for this listing, not a guarantee; rates vary by stream and task type

Dermatologist (MD) — Clinical Image Interpretation & AI Evaluation

What the work actually involves

What the screen is looking for

Logistics

Interview prep — 10 questions