About this role
We are seeking physicians to review, validate and design clinical tasks to help advance frontier AI models across a range of clinical scenarios and formats.
The project is designed to surface where models perform poorly in clinical scenarios such missed findings, incorrect clinical information, overconfidence or failure to seek context before answering.
You will work asynchronously on a dedicated platform, within your own specialty and day-to-day scope of practice.
This is a remote, flexible and short-term project.
KEY RESPONSIBILITIES
- Scenario Design: Design realistic clinical scenarios and instructions that mirror the work you carry out day-to-day in your specialty.
- Content Review: Assess clinical outputs for accuracy, flagging anything incorrect, outdated or can cause harm.
- Reference Answer Validation: Judge whether reference answers reflect what you would actually write in practice, including caveats, uncertainty and requests for further information.
- Gr...