What kind of job is data annotation?

Data annotation is remote contract work for AI companies. You look at data — a chatbot answer, a photo, an audio clip, a spreadsheet, a piece of code — and you label, correct, or rate it so a model can learn from your judgement. It sits somewhere between quality control, tutoring and freelance writing, and it is paid by the hour rather than by salary.

What a normal day looks like

You log into a platform dashboard, pick an available task queue, and work through items. A generalist shift might mean reading two AI answers to the same question and explaining which is better and why. A specialist shift might mean checking whether an AI's summary of a financial report or a clinical note is actually correct. Most tasks take 3–20 minutes and you can stop whenever you want.

Employment type and hours

• Contractor, not employee — you invoice hours; no holiday pay or benefits.
• Flexible hours — nights and weekends are fine on most platforms.
• Variable volume — work comes in project waves, which is why people join 2–3 platforms.
• Remote worldwide, though some language or US-only projects are restricted.

What it pays

Generalist text and image work pays roughly $15–$25/hr. Mid-level annotation and evaluation roles sit at $40–$60/hr. Domain experts — doctors, lawyers, accountants, engineers, PhDs — earn $70–$150/hr. You can see live ranges on our jobs board.

Skills that actually matter

Careful reading, following long instruction documents to the letter, clear written English, and patience. Coding or subject expertise raises your rate a lot, but is not needed to start. What gets people removed from projects is rushing, using AI to write their own answers, and ignoring guideline updates — see our assessment tips.

Who it suits — and who it doesn't

It fits students, parents, career changers, and professionals who want flexible extra income, plus experts who can bill high hourly rates part-time. It is a poor fit if you need a guaranteed weekly income, want a fixed salary and benefits, or dislike detailed written instructions.

Common questions

Is data annotation a full-time job?

Usually not. Almost all platforms hire independent contractors with flexible hours, so most people work part-time or stack two platforms to fill a full week.

Do you need a degree for data annotation?

No for generalist work — writing, reading comprehension and following instructions are enough. Specialist tracks paying $70/hr and up normally require a degree or professional experience.

Is data annotation work boring?

Generalist labeling can be repetitive. Evaluation and expert-review work is more varied because you judge model answers instead of drawing the same boxes.

How much can you make doing data annotation?

Generalist tasks pay about $15–$25/hr, mid-level annotation roles $40–$60/hr, and expert domain work $70–$150/hr.

Next: what data annotation is, whether it's legit, or how to get started.