Physician, Medical AI Evaluation
Posted 1hrs ago
Employment Information
Report this job
Job expired or something wrong with this job?
Job Description
US-licensed physicians evaluating clinical EHR vignettes and reasoning quality. Mindrift uses expert reviews to test and improve medical AI systems.
Responsibilities:
- Evaluate clinical EHR vignettes paired with questions, proposed answers, and distractor answers across diagnosis and treatment tasks
- Score clinical reasoning quality, vignette accuracy and completeness, answer correctness and gradeability, and trap quality
- Check whether reasoning chains reach answers from vignette facts alone
- Correct reasoning traces and write short rationales
- Participate in three independent blind reads and physician adjudication
- Complete qualification(s), join the project, submit tasks through the platform, and receive payment for approved tasks
Requirements:
- Medical degree (MD or DO)
- Active, unrestricted US medical license, verified at onboarding and re-checked monthly
- Board certification or completed residency training
- Direct patient care experience; attending-level preferred
- Broad, multi-system diagnostic and treatment experience in internal, family, emergency, or hospital medicine
- Strong clinical reasoning, including differential diagnosis, next-step management, and application of evidence-based guidelines
- Strong written English (C1+)
- US-licensed physician status
- Must not be a medical student, unlicensed medical graduate, non-physician clinician, or non-clinical healthcare professional
- Prior medical AI evaluation, clinical content or exam-question review, or medical annotation/QA experience is a plus
Benefits:
- Earnings per approved task: a fixed rate for each task, set in advance
- Fully self-scheduled: no minimum hours and no set schedule
- Remote work
- No patient contact and no malpractice exposure
- $50 one-time onboarding reward, paid once the first task on the project is approved
- Quality incentives: additional earnings for high-quality approved tasks
- Flexible participation around other commitments
- No time tracking or hour logging


