What the work actually is

You read studies and write down what a correct read looks like. Depending on the stream, that means labelling discrete findings against a defined schema — presence, location, laterality, severity, interval change — or authoring the full reference report a study should produce, or working a case with your diagnostic reasoning recorded step by step rather than compressed into an impression line. A second family of tasks points the other direction: reviewing AI-generated reads and flagging hallucinated findings, missed findings, and unsupported certainty; writing the grading criteria that separate an excellent read from a merely acceptable one; and constructing difficult cases designed to expose where model reasoning breaks down.

This is a shared expert pool. After onboarding you may be matched to any of several concurrent imaging workstreams and may move between them as priorities shift. Task length varies from roughly ten minutes for structured labelling to about an hour for full case authoring, and each stream carries its own throughput target. Nothing here touches live patient care.

What the screen looks for

  • Verifiable credentials first — MD or DO, completed diagnostic radiology residency, ABR certification or the equivalent in your jurisdiction (board-eligible with residency complete is accepted), an active unrestricted licence, and 2+ years reading independently post-residency.
  • Current, high-volume reading. Expect questions about your modality mix, weekly study volume, and how recently you were reading independently.
  • Fluency in structured reporting. BI-RADS, LI-RADS, Lung-RADS, TI-RADS and similar, as relevant to your subspecialty — including where the categories are genuinely ambiguous and how you resolve that.
  • Evaluation judgment. Whether you can articulate why a read is wrong, not just that it is, and whether you can write a criterion another radiologist would apply the same way.

Logistics

Fully remote and largely asynchronous, with a minimum of 15 hours per week and the ability to concentrate hours when a stream is time-boxed. Written and spoken English fluency is required, since much of the deliverable is prose. Fellowship training, U.S. licensure, prior annotation or dataset curation work, rubric design or resident assessment experience, and a linkable writing sample all help at matching time but are not gates. Pay of $200–400/hr reflects rates observed on this platform for radiology work; actual rates vary by stream, subspecialty, and task type.