How many AI evals jobs are open right now
As of 2026-10-20 there are 1147 open roles in AI evaluation, post-training, red teaming, alignment and human data across 170 companies. 566 of them publish a salary band, a median of $220k to $328k.
This page is rebuilt every morning from the career pages of 170 companies. It exists to be a citable number rather than a browsing experience. The roles themselves are on the board.
By discipline
By company
Where the work is. San Francisco, CA (476), New York, NY (229), London, UK (152), Mountain View, CA (70), Seattle, WA (60), Sunnyvale, CA (43), Washington, DC (40), US (35). 204 of 1147 roles are remote-friendly. 246 were posted in the last 30 days, counted across the 1051 listings that carry a posting date.
Free. One email on Thursdays.
Common questions
How many AI evals jobs are open right now?
1147 across 170 companies, as of 2026-10-20. The count is rebuilt every morning from company career pages.
What do AI evals jobs pay?
566 of the 1147 open roles publish a salary band. The median band runs $220k to $328k, and the full range across all published bands is $55k to $930k.
Which companies are hiring for AI evals?
Anthropic (97), OpenAI (54), Waymo (44), Scale AI (37), Prime Intellect (31), Amazon (31) are the largest by open role count.
How many AI evals jobs are remote?
204 of 1147, about 18%.
What counts as an AI evals job?
Evals and benchmarks, post-training and reinforcement learning, red teaming and safeguards, alignment and interpretability, and human data operations. General machine learning, infrastructure and gig annotation work are excluded.
Method
Collection. Once a day, from 170 companies' own public career endpoints. Nothing is submitted by employers and nothing is paid for inclusion.
What counts. Evals and benchmarks, post-training and reinforcement learning, red teaming and safeguards, alignment and interpretability, human data operations. A role must carry a signal in its title or department; matching on description text alone put lawyers and accountants on this board once and the rule exists because of it.
What is excluded. General machine learning and infrastructure roles, gig annotation and AI tutoring work, and anything posted more than 400 days ago.
Pay figures. Advertised annual USD base only, from the 566 of 680 pay-disclosing listings that give one. Hourly rates, non-USD bands and on-target figures are excluded rather than converted; those employers did publish pay, it is simply a different measure. Equity is not included, which matters most at the frontier labs where it is the larger half of the offer. Companies that publish pay skew toward the better-paying end, so read these as optimistic rather than as a market average.
Citing this page
These figures are current as of 2026-10-20 and change daily. If you are quoting them, quote the date with them.
Open roles in AI evaluation, post-training, red teaming, alignment and human data: 1147 across 170 companies (2026-10-20).
Stating pay in any form: 680 roles, 59%. Of those, advertising an annual USD base band: 566 roles, 49% of the board. Median advertised band $220k to $328k. Full range $55k to $930k.
Remote-friendly: 204 roles. Source: AI Evals Jobs, https://aievalsjobs.com/ai-evals-hiring-index/