Practising US physicians in primary care, internal medicine or hospital specialties score AI diagnoses and clinical reasoning for accuracy and safety. Suits licensed US MDs and DOs with recent post-residency practice who can commit 20+ hours a week.
Mercor is hiring experienced Medical Practitioners (MDs and DOs) to evaluate health AI models in partnership with a leading AI lab.
You will assess model outputs across diagnosis, clinical reasoning, and care workflows.
Key Responsibilities
\- Evaluate AI model outputs for clinical accuracy, safety, and alignment with medical standards
\- Score diagnoses, case reasoning, and care recommendations against expert judgment
\- Write structured feedback on clarity, safety, and clinical rigor
\- Flag failure modes that compromise patient safety
\- Collaborate asynchronously with AI researchers to improve model performance
Required Qualifications
\- MD or DO from an accredited medical school
\- Active, unrestricted US medical license
\- Residency complete
\- Currently practicing and seeing patients
\- 3+ years of post-residency experience in your current role strongly preferred
\- Practice in primary care, internal medicine, and hospital-based specialties (Primary Care, General Internal Medicine, Family Medicine, Emergency Medicine, Critical Care)
\- Sharp attention to detail and clear written communication
\- Located in the US
Project Timeline
\- Start Date: Immediate
\- Duration: 6 weeks
\- Commitment: Part-time, 20+ hours per week
Application & Onboarding
\- Upload your resume and application form
\- Complete a 25-minute conversational interview on your background, experience, and motivations
\- Credential verification: license status, residency completion, and current practice setting
\- Final roster placement is confirmed with the project team before your start date
\- Receive next steps and onboarding details within a few days
Apply today and put your clinical expertise to work shaping the health models clinicians and patients will rely on.
Listing from Mercor's job board. Details can change; the platform's page is the final word.
