Methodology

• remowork.life. Most listings are provided by remowork.life, which gathers remote AI training jobs from company and platform job boards. We read them through a partner connection once a day. We show each job as provided and link to the employer's own application.
• Employers. Employers can post listings here. Each is reviewed before it goes live and runs for 30 days.

Listings refresh once a day. A job that no longer appears in a complete refresh is marked closed, and closed jobs stay viewable with a notice but are left out of counts, pay statistics, and lists. Company profiles refresh at the same time.

• Only open listings that post a pay range in US dollars are counted.
• The midpoint of the range becomes an hourly rate: daily pay is divided by 8, weekly by 40, monthly by 173, and yearly by 2,080 hours.
• Per-task and per-project pay is left out, because it cannot be converted to an hourly rate fairly.
• Rates below $1 or above $1,000 an hour are dropped as probable data-entry errors.
• Each group (a field, a platform, a country) shows its minimum, 25th percentile, median, 75th percentile, and maximum, plus how many listings it is based on. The “typical range” is the 25th to 75th percentile.

These figures describe what employers advertise, not what workers earn. See the disclaimer. The datasets and the same method are on the data page.

• Job counts are open jobs only. A job that needs several languages counts once for each language.
• A country's count is the jobs a person there can apply to: jobs that name the country, plus jobs open to applicants everywhere, minus jobs that exclude it.
• Lists such as “no experience required” or “pays weekly” are built from wording in the listing and say so. They show what employers state, and the employer's own page is the authority.

We ask search engines to index a job page only when the job is open, has a description of at least 500 characters, and has a way to apply. Closed and near-empty listings stay reachable but are marked so they do not clutter search results.

We review employer listings before publishing and remove any listing, from any source, that asks applicants to pay, uses a cheque or gift-card scheme, or otherwise misleads workers. Anyone can report a listing from its page or at abuse@aitrainingjobs.app. See how to tell real work from a scam.

Guides are written to answer common questions and link to the live pay data instead of repeating numbers that go out of date. Some are drafted with AI writing tools, as the disclosures explain. Figures on the site come from code, not from AI.

Found an error? Write to support@aitrainingjobs.app with the page address. We fix confirmed mistakes and update the page.