Nephrologist, Paid AI Evaluation Project | $200–400/hr, Starts Now

Nephrologist, Paid AI Evaluation Project | $200–400/hr, Starts Now
This is a paid, part-time project, not a job. It starts immediately, you choose your hours, and if it goes well there will be more projects after it. Bring your clinical judgment to the cases medical AI gets wrong, work directly with our team, and help shape a research paper.
The project
Archangel Health has an active project that needs practicing nephrologists now. You’ll review difficult, fully de-identified nephrology cases, evaluate AI responses, and write detailed rubrics that explain what good clinical reasoning requires. When a case has more than one reasonable answer, you’ll help distinguish acceptable approaches from unsafe decisions. Your work will also support a research paper we’re developing with physicians.
We’re a Y Combinator-backed team of Berkeley and OpenAI engineers building tools to make medical AI safe and capable.
Pay and hours
$200 to $400 per hour depending on experience, paid directly to you. The project starts as soon as you’re ready. Hours are entirely up to you: 1 hour or 20, with no minimum. Every hour is paid, including training. You’ll work directly with our team, with guidance at every step.
This is project-based work with a defined scope. It is not a full-time or ongoing employment position. Nephrologists who do well on this project will be first in line for future ones.
Who this is for
Practicing nephrologists who are board-certified or hold the equivalent credential, are fluent in English, and work in US or nearby US time zones. We value sound clinical judgment and the ability to explain difficult decisions clearly. Prior AI evaluation, data-labeling or rubric-writing work is a strong plus.
Why this matters
A medical AI response can sound convincing and still miss what matters for a patient. Nephrologists can spot those gaps and explain how a safer decision should be made. Your clinical reasoning will help us build more demanding evaluations and identify where these systems need to improve.

Work arrangement
Yes