Skip to content

Research report ยท September 2026 edition

Where AI Labs Are Pushing: The Expertise Being Hired, Week by Week

From the week of 7 Sep 2026 to the week of 21 Sep 2026, too few fields, languages or kinds of work had enough open roles on the 0 platforms tracked throughout to compare their growth.

Read the month's story: The AI Training Job Market in September 2026

Data window
1 Sep 2026 to 30 Sep 2026
30 days, UTC
Frozen
11 Nov 2026, 14:35 UTC
These figures will not change
Source
The public job boards of the AI training platforms this site tracks, read every hour. Fields and languages are read from role titles. Listings only.
Method
Version 3. Definitions
CC BY 4.0, how to cite

Every figure on this page as CSV and JSON, sent to your work email.

Key findings

  1. 1

    On the 0 platforms tracked throughout, from the week of 7 Sep 2026 to the week of 21 Sep 2026, 0 groups began to get new listings in the last 1 weeks after none in the first 2, and 0 got none in the last 1 after some before; no group appeared in or vanished from the open roles.

    Figure 7

Figure 1

Fastest-growing expert fields

Which professions and disciplines are AI training platforms hiring more of?

Expert fieldChange in open rolesOpen, week of 7 SepOpen, week of 21 SepNew listings, last 1 week vs 1 before
All roles
n/a
00n/a

Sample. Roles on the 0 platforms tracked throughout, over the 3 complete ISO weeks from the week of 7 Sep 2026 to the week of 21 Sep 2026; 0 expert fields named in titles.

Open roles: the average number listed on the days of the week (00:00 UTC each day), on the 0 platforms tracked throughout. A platform first tracked later (0 in this window) is left out of every week, so its arrival never reads as growth.

Change: from the first to the last complete week, a rate (open roles in the last week over the first). No row with any open roles in the first week is a small sample: the rows differ so much that each reads on its own from its first.

New listings, last 1 week vs 1 before: new listings a week over the later half of the complete weeks against the earlier half (1 on each side); shown when the weeks before hold at least 1 new listings, the weight of the expert fields' average pace fitted on their own counts (below it the average weighs more than the group's own). A new listing is a role first seen that week on a platform past its first week in this dataset.

A role counts for an expert field when its title names it (whole words, the vocabulary of the expert fields report), and for each one it names.

Source. AITraining.jobs Research, Where AI Labs Are Pushing: The Expertise Being Hired, Week by Week, September 2026 edition.

Figure 2

Fastest-growing languages

Which languages are AI training platforms hiring more of?

LanguageChange in open rolesOpen, week of 7 SepOpen, week of 21 SepNew listings, last 1 week vs 1 before
All roles
n/a
00n/a

Sample. Roles on the 0 platforms tracked throughout, over the 3 complete ISO weeks from the week of 7 Sep 2026 to the week of 21 Sep 2026; 0 languages named in titles.

Open roles: the average number listed on the days of the week (00:00 UTC each day), on the 0 platforms tracked throughout. A platform first tracked later (0 in this window) is left out of every week, so its arrival never reads as growth.

Change: from the first to the last complete week, a rate (open roles in the last week over the first). No row with any open roles in the first week is a small sample: the rows differ so much that each reads on its own from its first.

New listings, last 1 week vs 1 before: new listings a week over the later half of the complete weeks against the earlier half (1 on each side); shown when the weeks before hold at least 1 new listings, the weight of the languages' average pace fitted on their own counts (below it the average weighs more than the group's own). A new listing is a role first seen that week on a platform past its first week in this dataset.

A role counts for a language when its title names it (whole words, the vocabulary of the languages report), and for each one it names.

Source. AITraining.jobs Research, Where AI Labs Are Pushing: The Expertise Being Hired, Week by Week, September 2026 edition.

Figure 3

Fastest-growing kinds of work

Which kinds of AI training work are platforms hiring more of?

Kind of workChange in open rolesOpen, week of 7 SepOpen, week of 21 SepNew listings, last 1 week vs 1 before
All roles
n/a
00n/a

Sample. Roles on the 0 platforms tracked throughout, over the 3 complete ISO weeks from the week of 7 Sep 2026 to the week of 21 Sep 2026; 0 kinds of work.

Open roles: the average number listed on the days of the week (00:00 UTC each day), on the 0 platforms tracked throughout. A platform first tracked later (0 in this window) is left out of every week, so its arrival never reads as growth.

Change: from the first to the last complete week, a rate (open roles in the last week over the first). No row with any open roles in the first week is a small sample: the rows differ so much that each reads on its own from its first.

New listings, last 1 week vs 1 before: new listings a week over the later half of the complete weeks against the earlier half (1 on each side); shown when the weeks before hold at least 1 new listings, the weight of the kinds of work' average pace fitted on their own counts (below it the average weighs more than the group's own). A new listing is a role first seen that week on a platform past its first week in this dataset.

Every role is classified into one kind of work.

Source. AITraining.jobs Research, Where AI Labs Are Pushing: The Expertise Being Hired, Week by Week, September 2026 edition.

Figure 4

New listings per week, by expert field

Week by week, how many new roles do platforms list in each of the largest expert fields?

No data in this window.

Sample. New listings on the 0 platforms tracked throughout, by ISO week (Monday to Sunday, UTC), over the 3 complete ISO weeks from the week of 7 Sep 2026 to the week of 21 Sep 2026: the 0 expert fields with the most.

A column is the week starting on the Monday it names. Each platform's first week in this dataset (its back catalogue) is not counted as new listings.

One platform posting a batch shows as a single dark week; the growth table says how many platforms listed new roles in each group.

A row is marked as a small sample when its movement from week to week does not read on its own: the acceleration of the growth table, shown when the weeks before hold at least 1 new listings, the weight of the expert fields' average pace fitted on their own counts (below it the average weighs more than the group's own).

Source. AITraining.jobs Research, Where AI Labs Are Pushing: The Expertise Being Hired, Week by Week, September 2026 edition.

Figure 5

New listings per week, by language

Week by week, how many new roles do platforms list for each of the most-hired languages?

No data in this window.

Sample. New listings on the 0 platforms tracked throughout, by ISO week (Monday to Sunday, UTC), over the 3 complete ISO weeks from the week of 7 Sep 2026 to the week of 21 Sep 2026: the 0 languages with the most.

A column is the week starting on the Monday it names. Each platform's first week in this dataset (its back catalogue) is not counted as new listings.

One platform posting a batch shows as a single dark week; the growth table says how many platforms listed new roles in each group.

A row is marked as a small sample when its movement from week to week does not read on its own: the acceleration of the growth table, shown when the weeks before hold at least 1 new listings, the weight of the languages' average pace fitted on their own counts (below it the average weighs more than the group's own).

Source. AITraining.jobs Research, Where AI Labs Are Pushing: The Expertise Being Hired, Week by Week, September 2026 edition.

Figure 6

New listings per week, by kind of work

Week by week, how many new roles do platforms list in each of the largest kinds of AI training work?

No data in this window.

Sample. New listings on the 0 platforms tracked throughout, by ISO week (Monday to Sunday, UTC), over the 3 complete ISO weeks from the week of 7 Sep 2026 to the week of 21 Sep 2026: the 0 kinds of work with the most.

A column is the week starting on the Monday it names. Each platform's first week in this dataset (its back catalogue) is not counted as new listings.

One platform posting a batch shows as a single dark week; the growth table says how many platforms listed new roles in each group.

A row is marked as a small sample when its movement from week to week does not read on its own: the acceleration of the growth table, shown when the weeks before hold at least 1 new listings, the weight of the kinds of work' average pace fitted on their own counts (below it the average weighs more than the group's own).

Source. AITraining.jobs Research, Where AI Labs Are Pushing: The Expertise Being Hired, Week by Week, September 2026 edition.

Figure 7

Groups that appeared or vanished

Which fields, languages and kinds of work started or stopped being hired over the weeks?

No data in this window.

Sample. None of the expert fields, languages or kinds of work on the 0 platforms tracked throughout appeared, vanished, began or stopped getting new listings over the 3 complete ISO weeks from the week of 7 Sep 2026 to the week of 21 Sep 2026.

Appeared: no role open on any day of the first week, and at least one open on average in the last. Vanished: the reverse.

New listings began: no new listing in the first 2 weeks, at least one in the last 1. Stopped: the reverse. Many platforms keep roles listed for months, so a group can stop getting new listings while its open roles stay up.

A row is marked as a small sample when the side that moved (open roles for appeared or vanished, new listings otherwise) is smaller than the weight of its family's average change in that measure, fitted on the family's own counts (the change and acceleration notes of the growth tables): one or two postings can make a group appear or vanish.

Source. AITraining.jobs Research, Where AI Labs Are Pushing: The Expertise Being Hired, Week by Week, September 2026 edition.

For labs, data vendors and researchers

Buy the data behind this report

The figures above are a summary. Behind them are records we collect ourselves, every day:

  • 0 roles tracked, each with its history

    Every role on 0 job boards, read since n/a (every hour since 18 Oct 2026): its pay and place, and the first and last day it was open.

  • 0 apply clicks in 28 days

    Which roles people actually apply to, by field, pay and kind of work. It happens on this site, so nobody else can count it.

  • 39 lab and vendor hiring boards

    The job boards of 25 AI labs and 14 human-data vendors, read every day: who is hiring for human data, and who is hiring to buy it.

What you can buy

  • The records

    The rows behind any figure, with their history, not only the totals.

  • A regular feed

    The same figures every week or month, as files.

  • A custom cut or benchmark

    Your countries, languages or fields, or your own platform against the market.

Building AI and need the experts behind this data? (you contract them yourself) or (we deliver the expert work).

Investing in AI-data companies? Hiring radar is these records role by role, every day, as files.

Method and definitions

The public job boards of the AI training platforms this site tracks, read every hour. Fields and languages are read from role titles. Listings only. Every figure is computed by one database query over the window shown, the same query for the live report and for every edition. Each figure states its sample beneath it.

Fixed panel
The platforms tracked for at least a full week before the first week counted. A platform added later never reads as growth, and neither do the old roles a newly tracked board surfaces in its first days.
Open roles in a week
The average number of the group's roles listed on the week's days (00:00 UTC each day), on the fixed panel.
Change
Open roles in the last complete week against the first, in percent. A group's change reads on its own when its open roles in the first week reach the weight of its family's average change, fitted on the family's own counts (below it the average weighs more than the group's own); other groups are marked as small samples and never ranked.
New listing
A role that first appeared during the week on a platform already tracked before that week. A platform's first week in this dataset is its back catalogue and is not counted as new. When this site changes how it reads a platform's board (a cap lifted, a page kind or department read too, a relevance rule changed), the roles that came into view or left it from that change to the end of its week are this site's reading, not the platform's hiring: they are not new listings or removals, and changes in open roles are measured like for like across it. Each platform's report says what changed.
Acceleration
New listings a week over the last four complete weeks against the four before, in percent (two against two in an edition month with four complete weeks). Shown when the weeks before hold at least the weight of the family's average pace in new listings, fitted on the family's own counts.
Appeared, vanished, began, stopped
Appeared: no role open on any day of the first week, at least one on average in the last; vanished, the reverse. Began: no new listing before the last weeks of the acceleration comparison, some in them; stopped, the reverse.
Expert field and language
Named in a role's title, matched as whole words with the vocabularies of the expert fields and languages reports. A role naming two counts for both.
Kind of work
Each role is classified from its title and description into one of the twelve kinds of AI training work this site tracks. Every kind links to its own page.
Window
Live: the last twelve complete ISO weeks (Monday to Sunday, UTC), from the week of 3 Aug 2026 at the earliest. Edition: the complete ISO weeks of one calendar month.

What this data cannot tell you

  • Platforms recruit for their clients' projects, so hiring is a sign of where labs buy data, not a statement by any lab. Which lab is behind a role is not published.
  • One platform posting a batch moves a group's numbers; the growth tables say how many platforms listed new roles in each group.
  • A listing is not a hire. Some platforms keep evergreen roles open for months, and some post one listing per language or country.
  • Generic titles hide the field or language of many roles, so a group's counts are a floor.
  • Platforms that recruit without a public job board are not counted, and platforms added to the dataset after the window opens are left out of it.
  • Before 3 Aug 2026 this dataset's feeds were still being added and tuned, so the series does not reach back further.

How the job feeds are read and roles are classified: site methodology. How fresh, complete and consistent the data is, checked every day: data quality and coverage.

Editions

The live report covers the last 12 complete weeks (from 3 Aug 2026 at the earliest) and moves on every Monday. Each complete month is also published as an edition, frozen two days after the month ends, so a figure you cite from an edition never moves.

Cite and reuse

Free to quote, chart and republish under CC BY 4.0 (creativecommons.org/licenses/by/4.0) with a link to this page. This edition is frozen, so its figures will read the same later.

Citation

AITraining.jobs (2026). Where AI Labs Are Pushing: The Expertise Being Hired, Week by Week, September 2026 edition. https://cyberjobs.cc/research/where-labs-are-pushing/2026-09

BibTeX

@misc{aitrainingjobs_where_labs_are_pushing_2026_09,
  author = {{AITraining.jobs}},
  title = {Where AI Labs Are Pushing: The Expertise Being Hired, Week by Week, September 2026 edition},
  year = {2026},
  howpublished = {\url{https://cyberjobs.cc/research/where-labs-are-pushing/2026-09}},
}
Data: every figure as CSV (one line per value) and JSON, sent to your work email.Pay and open-roles data

More research

All 81 research reports