Your judgment trains the models. Get paid for it — and get known for it.
Evaluate AI answers, rank responses, annotate data, review generated code, and find model failures. Remote, flexible, priced per verified item — with a public reputation that follows you.
Paid per verified item
No hourly race. Domain experts (math, law, medicine, code) earn premiums. Payouts via PayPal, minimum $10.
A profile, not a queue
perbug.com/@you shows your skills, specializations, accuracy, and history — the thing AI companies search to recruit.
Qualifications unlock pay
Pass a task-type test once; higher accuracy and tier unlock private, higher-paying programs.
You own your privacy
Show your credibility; keep earnings and client names private by default.
Real variety
Response evaluation, RLHF, annotation, coding eval, search eval, domain tasks, red teaming, rewrites.
Fair quality control
Gold items and consensus keep scoring honest — and protect good workers from bad clients.
Measured across 4,281 completed evaluations, blind quality checks, domain tests, consistency, and client feedback — not self-reported.