On 14 October 2026, AI tutor roles listed a median of $40 an hour and AI evaluator roles a median of $18 an hour. An AI tutor writes and corrects the answers an AI model learns from, and an AI evaluator rates answers the model has already given, or search results, against a set of rules. The figures come from the 20 open roles on this site with "tutor" in the title and the 91 with "evaluator" or "evaluation".
Both lists are live: AI tutor jobs and AI evaluator jobs. The numbers below are worked out again every time the lists are read.
AI tutor jobs
The tutor roles open today are at Meridial and xAI. xAI, the lab that makes the Grok models, calls its trainers "AI Tutors" and posts one role per language or subject on its own job board. A typical title is "AI Tutor" followed by a language or a subject.
| Platform | Open roles | With an hourly rate | Median listed rate | Middle half |
|---|---|---|---|---|
| xAI | 19 | 19 | $40 | $40 to $55 |
| Meridial | 1 | 1 | too few rates |
20 of the 20 tutor roles post an hourly rate. The middle half runs from $40 to $55.
What the work asks for, going by the listings: strong writing in the language of the role, the habit of checking facts, and comfort with detailed written guidelines. Subject roles add a degree or work experience in the subject. The xAI page has every open xAI role and how its hiring works.
AI evaluator jobs
Evaluator roles are open at 8 platforms: Appen, Innodata, Meridial, micro1, Mindrift, RWS TrainAI, Terac and Welocalize. 42 of the 91 are search engine evaluator roles. In that work you judge whether the results a search engine shows match what the person was looking for, in your language and for your country.
| Platform | Open roles | With an hourly rate | Median listed rate | Middle half |
|---|---|---|---|---|
| Terac | 10 | 9 | $65 | $50 to $85 |
| micro1 | 6 | 6 | $40 | $35 to $63.75 |
| Meridial | 7 | 7 | $35.50 | $35.50 |
| RWS TrainAI | 55 | 52 | $14 | $6.75 to $18 |
| Innodata | 5 | 4 | too few rates | |
| Appen | 4 | 0 | too few rates | |
| Mindrift | 3 | 3 | too few rates | |
| Welocalize | 1 | 1 | too few rates |
45 evaluator listings give only a top rate ("up to"), so the median here is partly a median of ceilings. Listed rates run from $4 to $110 an hour. Search evaluator roles are posted once per country, each with its own rate, so the same work can pay very different amounts in different places.
The rest are evaluator roles tied to a field, such as medicine or coding, or general quality checks of model answers.
Tutor or evaluator: the difference in practice
- Tutors write. You produce questions, model answers and corrections, so the job rewards writing speed and care.
- Evaluators judge. You read and score, usually against a long rule book, so the job rewards consistency.
- Language matters for both. Most of these roles ask for one language. The roles by language page lists what is open in each.
Most platforms test you before the first paid task. The interview prep pages have practice questions by type of role.
Sources
- This site's list of open roles (jobs.json), read on 14 October 2026.
- xAI job board (Greenhouse), read on 14 October 2026.
- RWS job board (Lever), read on 14 October 2026.