Get paid. Qualify for AI subcontractor work.
Jobpeak is the platform for people who want to do AI support work — reviewing model outputs, classifying content, evaluating responses. One test. no CV. no interview. Bi-weekly USD payments at rates up to $100/hr once qualified.
Jobpeak qualifies independent subcontractors for paid AI support work — the kind of work that powers the AI models you already use. Tasks include reviewing AI-generated answers, spotting errors, classifying content, and evaluating model responses for accuracy and tone. We don't screen by CV. We don't run interviews. The only thing that matters is whether you can do the work, and we test for that directly with our domain-specific Cognition Test. Pass it, get qualified, start earning.
Built for the way independent work actually pays.
Bi-Weekly USD payments via Stripe or PayPal. Four registered entities across the US, Canada, and Australia and New Zealand. Transparent tier structure with rates published up front.
What you can earn
Rates by tier.
Pay scales with the difficulty of the work and the rigour of the qualification. Rates published up front.
BASIC
Up to $30/hr
Generalist work — reviewing AI outputs across everyday subjects, content classification, and basic annotation. The entry route into the platform.
STEM
Up to $40/hr
Specialist work in mathematics, physics, chemistry and biology. Requires passing the Cognition Test in your chosen science.
CODING
Up to $100/hr
Code review, debugging AI-generated code, and evaluating model outputs across programming languages. The highest live rate ceiling on the platform.
PROFESSIONAL
Up to $50/hr
Finance and accounting are live. Law and medicine are in development register your interest to be notified when those open.
Three things every subcontractor needs.
We don't test for credentials. We test for the dispositions that actually predict good AI review work.
Domain knowledge
40 questions drawn from a 150-question bank in your chosen specialty. Calibrated across easy, medium and hard. Including a section on reviewing AI outputs in your domain because that's what the work actually is.
Calibrated judgement
20 behavioural questions sampled proportionally from six trait categories: calibration, instruction-following, ambiguity tolerance, attention under fatigue, bias awareness, and communication. The traits that separate a good reviewer from a credentialed one.
Self-awareness
Bias awareness, communication clarity, and the willingness to flag what you don't know.
