Licensed Physician (MD/DO) - Freelance AI Trainer

Mindrift•San Antonio, TX

5d•Remote

About The Position

Mindrift is seeking actively practicing physicians with an active, unrestricted US medical license to evaluate clinical EHR vignettes for a medical AI evaluation project, with a primary focus on general and internal medicine. Contributors review written clinical cases and help test, evaluate, and improve AI systems. The work involves no patient contact and is project-based, with estimated contributions of around 10–20 hours per week during active project phases. Contributors choose when and how much to work, subject to project timelines and acceptance criteria. Physicians assess clinical benchmark items across multiple body systems, evaluating vignette accuracy and completeness, whether cases reveal their answers, answer correctness and gradeability, and the quality of distractor trap answers. They also assess clinical reasoning chains, correct reasoning traces, and write short rationales. Work proceeds through three independent blind reads followed by physician adjudication. Compensation is a flat rate per approved task, estimated to be equivalent to about $100 per hour based on task estimates. QA and adjudication tasks are paid at a higher rate. A one-time $50 onboarding reward is paid after the first production task is approved, and additional incentives may be available for high-quality approved tasks. Participation and workload are project-based; the estimated weekly hours are not a minimum or guaranteed workload.

Requirements

Medical degree (MD or DO).
Active, unrestricted US medical license, verified at onboarding and re-checked monthly.
Board certification or completed residency training.
Recent direct patient care experience; attending-level experience is preferred.
Broad, multi-system diagnostic and treatment experience in internal, family, emergency, or hospital medicine.
Strong clinical reasoning skills, including differential diagnosis, next-step management, and application of evidence-based guidelines.
Strong written English proficiency (C1+).

Nice To Haves

Prior experience in medical AI evaluation, clinical content or exam-question review, or medical annotation/QA.

Responsibilities

Evaluate clinical EHR vignettes paired with questions, proposed answers, and distractor trap answers across diagnosis and treatment tasks and multiple body systems.
Score benchmark items for vignette accuracy and completeness, answer leakage, answer correctness and gradeability, and trap quality.
Check whether reasoning chains reach answers using vignette facts alone and assess the correctness of reasoning traces.
Write short rationales for evaluations.
Participate in three independent blind reads followed by physician adjudication.
Submit tasks within project timelines and meet the stated acceptance criteria.