Frequently asked questions
Everything you might want to know about working with Aiyana — as a reviewer or as a research partner.
Working with Aiyana
Do I need an AI or ML background?
No. We hire domain experts — clinicians, lawyers, writers, engineers, researchers, translators. We train you on the AI-side of the work; your expertise is what we need.
How much can I earn?
Most experts earn between $25 and $55/hr depending on domain, seniority, and project. Senior reviewers and project leads earn meaningfully more. Pay is in USD, weekly.
Is the work really remote?
Yes — 100% remote, fully async. You need a laptop, a stable internet connection, and a quiet workspace. That's it.
How many hours can I work?
You set your own hours. Some experts work 5 hours/week, some work 30+. We have a soft floor of 5/week to keep your calibration warm and a ceiling of 40/week to avoid burnout.
Application & onboarding
How long does the application take?
Around 8 minutes. We ask about your background, expertise, and work preferences — no writing samples or essays.
What does the assessment look like?
A short paid calibration task in your domain — usually 30–60 minutes. If it goes well, we schedule a brief async interview and onboard you.
How quickly do I hear back?
Most candidates hear back within 5 business days of submitting. If your domain is in active demand, sometimes within 24 hours.
Do I need to send a CV?
No CV required. We use a structured application that captures the signal we actually care about. If you'd like to attach LinkedIn or a portfolio, you can.
Payments & taxes
How and when do I get paid?
Weekly, every Friday, via Wise or Deel. We support 60+ countries and most major currencies (settlement in USD).
Am I an employee or contractor?
Independent contractor in most jurisdictions. You're responsible for your own taxes and we provide year-end summaries.
Are there bonuses?
Yes. Top reviewers (by quality and consistency) earn quarterly performance bonuses and are first in line for senior reviewer roles.
About the work
What kinds of projects will I work on?
Evaluating model outputs, writing reference answers, designing rubrics, red-teaming for safety, and grading agent trajectories. Specific projects depend on your domain and what we're running.
Is the work interesting?
Most of our experts say yes. The closest analog is peer review — the work is genuinely intellectual and you can see your impact in the next model release.
Will I see my work in deployed products?
Not directly attributed (NDAs apply), but yes — the models you train are used by millions of people every day.
What happens if I make a mistake?
Mistakes are normal — we flag them through gold checks and reviewer feedback. We focus on calibration, not punishment. Consistent quality issues trigger coaching first, not termination.
Privacy & ethics
How do you handle my data?
GDPR/CCPA-aligned. Your personal data is encrypted, used only for vetting and payroll, and you can request deletion at any time.
What about confidentiality?
Every cohort signs an NDA with the client. Project content is strictly confidential — don't share screenshots, prompts, or outputs anywhere.
Do you work on harmful AI?
No. We explicitly decline projects involving CBRN uplift, mass surveillance, autonomous weapons, or non-consensual content. Where we do safety work, it's exclusively to reduce harm.