A nurse making $40 an hour at the bedside can open a job board today and find clinical AI evaluation roles posted at $80 to $120 an hour for the same judgment. That gap is the whole business.

Lawyers and compliance professionals see the same split. Rating a model's analysis, catching the case citation it invented, writing the answer it should have given: contract review, legal research, statutory reasoning, AML and sanctions work, posted at $90 to $150 an hour. Bar admission, a practising certificate, a paralegal qualification or a compliance background all get looked at.

Senior software engineers see the largest numbers. Reviewing AI-written code, scoring an agent's whole trajectory, writing the reference implementation: $150 to $210 an hour for ten years or more of experience, and it is the best paid of the three right now. It is also a clean bridge between roles, with no notice period to serve.

If you hold one of those credentials, the work is not "learn AI." It is packaging the judgment you already get paid for, and then sitting in a room with other people who do the same work.

That is what Ground Truth is. One guide per profession: clinical, legal and compliance, and senior software engineering. A winners chat with people who hold your credential. Optional weekly listing drops if you want the board brought to you.