QA Auditor — AI Training Data

You will audit completed annotations produced by other team members, checking each one against the project rubric and flagging errors. This is second-pass work: someone else has already labeled the data, and your job is to verify that the labels are correct, consistent, and complete. On a typical day you will review 200-400 items, writing short justifications for every rejection. The work requires sharp attention and the ability to hold a full annotation schema in your head while scanning at speed.

You should have a track record of careful, systematic work. Backgrounds in copy editing, test engineering, clinical data review, or research assistance translate well. We will train you on our specific tools and rubrics, but we cannot teach the underlying mindset: you either notice when something is slightly off, or you do not.

Shifts are flexible. Some reviewers prefer to batch their work into two or three longer sessions per week; others spread it out daily. We are agnostic about schedule as long as turnaround targets are met. A weekly 30-minute sync call with the quality team is the only fixed calendar item.