Key findings
- 1
Mindrift listed 90 open roles on 31 Oct 2026, 2.1% of the 4,381 open across the 62 platforms AITraining.jobs reads, the 16th most.
- 2
As this site read it, Mindrift went from 0 open roles on 4 Oct 2026 to 90 on 31 Oct 2026, but that count changed with our reading of its board as well as with its hiring: roles came into and left view when this site changed how it reads the board on 14 Oct 2026 and 22 Oct 2026. Measured on the roles that did not move with it, like for like its open roles held level.
- 3
41.1% of its open roles are in Law, Medicine & Finance, where Mindrift holds 2.4% of all open roles, the 14th largest share of 46 platforms; it lists 3 of the 12 kinds of work with open roles.
- 4
On 31 Oct 2026, 66.7% of its open roles stated an hourly rate. Their 19 distinct advertised rates (41 more are only an advertised maximum, "up to $X/hr", left out) have a median of $100/hr and typical pay of $60 to $100/hr, against a median of $40/hr across other platforms' rates.
- 5
In Law, Medicine & Finance, the one kind of work where both Mindrift and other platforms state enough rates to compare, its median is $100/hr against $70/hr (+42.9%).
- 6
None of its 90 open roles are open to applicants in any country (18.9% on other platforms); all are limited to listed countries, most often the United States (named by 54.4%).
- 7
From 1 Oct 2026 to 31 Oct 2026, Mindrift posted 232 new listings and removed 146 (listed a median of 7.3 days before removal, against 36.5 days on other platforms); its open roles went from 0 to 90 as read, and held level like for like, across the change in how this site reads its board on 14 Oct 2026 and 22 Oct 2026.
- 8
Of the 90 roles Mindrift lists with a full description, 45.6% mention a PhD or master's degree, 44.4% a professional credential and 0.0% an assessment or test (other platforms with descriptions: 24.7%, 16.0% and 20.4%).
Figure 1
Mindrift against the other platforms
How does Mindrift compare with the rest of the market on size, pay, reach and turnover?
| Measure | Mindrift | Other platforms | Its rank (1 = highest) |
|---|---|---|---|
| Open roles on 31 Oct 2026 | 90 | 4,291 | 16th of 62 |
| Share of all open roles | 2.1% | 97.9% | 16th of 62 |
| New listings in the window | 232 | 1,406 | 3rd of 62 |
| Kinds of work listed | 3 | 12 | joint 36th of 62 |
| Open roles stating an hourly rate | 66.7% | 81.7% | joint 37th of 62 |
| Median advertised pay | $100/hr | $40/hr | joint 6th of 37 |
| Open roles accepting any country | 0.0% | 18.9% | joint 24th of 62 |
| Median days listed, removed roles | 7.3 days | 36.5 days | 25th of 26 |
Sample. 90 roles Mindrift listed open at the end of 31 Oct 2026, beside the 4,291 open on the other 61 platforms AITraining.jobs reads; listings from 1 Oct 2026 to 31 Oct 2026.
Other platforms: every other platform read over the same window, pooled (counts are summed, medians are taken over all their distinct rates together).
Pay is measured as on the job board: one sample per distinct advertised rate (the middle of a stated range; a role posted again at the same rate counts once), and an advertised maximum ("up to $X/hr") is a ceiling, not a rate, so it is counted and left out. Typical pay is the 25th percentile of what roles pay (the middle of each advertised range) to the 90th-percentile advertised ceiling; it and the median are rounded to $5 ($1 under $20), as on every list.
Rank among the 62 platforms with open roles. A share is ranked where the platform's open roles outweigh the average of the platforms it leans on, the weight fitted on every platform's own count (here 0.8 open roles for stating a rate, 0.4 open roles for accepting any country); a pay median needs 3 distinct advertised rates and days listed 3 removed roles with a known listing date. A platform below one is not ranked on it. Days listed rank longest first.
Listings only: what the platform's public job board shows.
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 2
Open roles at Mindrift, week by week
Is Mindrift listing more AI training roles or fewer?
| Week | Open roles, like for like (first week = 100) | Open at the end of the week, as read | New listings |
|---|---|---|---|
| 1 to 4 OctPartial week | n/a | 0 | 0 |
| 5 to 11 Oct | 100.0 | 28 | 28 |
| 12 to 18 Oct (our reading changed)Partial week | 246.4 | 296 | n/a |
| 19 to 25 Oct (our reading changed)Partial week | 246.6 | 56 | 46 |
| 26 to 31 OctPartial week | 396.3 | 90 | 97 |
Sample. Listings by Mindrift, by ISO week (Monday to Sunday, UTC), 1 Oct 2026 to 31 Oct 2026; 5 weeks tracked.
Open roles are counted at the end of each week (Sunday, UTC), or at the end of the window for a week it cuts. New listings leave out roles that came with the platform's back catalogue.
Like for like: open roles against the first week with open roles (100), measured across any change in what AITraining.jobs reads of the board on the roles that neither came into view nor left it with the change, so our reading never reads as hiring.
On 14 Oct 2026 our reading of Mindrift changed: this site began reading its whole Workable board (950 rows, 399 distinct jobs) instead of its first 300 rows, its new postings judged over the next reads. From then to the end of that week (18 Oct 2026) the roles that came into view are counted as already listed, not as new listings and the roles that left view are not counted as removals, and its open roles are compared like for like across the change.
On 22 Oct 2026 our reading of Mindrift changed: this site began reading each posting's own text in its relevance check, and left out its three-day paid language studies in New York, which are not AI training work. From then to the end of that week (25 Oct 2026) the roles that left view are not counted as removals, and its open roles are compared like for like across the change.
A week the window cuts is labelled with the days it covers and marked as partial.
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 3
What Mindrift hires for
Which kinds of AI training work does Mindrift list, and how much of each market does it hold?
| Kind of work | Open roles | Share of its roles | Its share of the kind's open roles |
|---|---|---|---|
| Law, Medicine & Finance | 37 | 41.1% | 2.4% |
| Robotics & Physical-AI Data | 30 | 33.3% | 30.0% |
| Agentic Tasks & RL Environments | 23 | 25.6% | 18.5% |
Sample. 90 roles Mindrift listed open at the end of 31 Oct 2026; market shares are of the 4,381 roles open on all 62 platforms that day.
Kinds of work are this site's twelve categories of AI training work, each role classified from its title and description.
Its rank in the kind is among the platforms with open roles in it. A median needs at least 3 distinct advertised rates, as on the job board.
Pay is measured as on the job board: one sample per distinct advertised rate (the middle of a stated range; a role posted again at the same rate counts once), and an advertised maximum ("up to $X/hr") is a ceiling, not a rate, so it is counted and left out. Typical pay is the 25th percentile of what roles pay (the middle of each advertised range) to the 90th-percentile advertised ceiling; it and the median are rounded to $5 ($1 under $20), as on every list.
Rows with fewer than 7.5 open roles in the kind are marked as a small sample and not ranked: below that the average of the rows weighs more than a row's own open roles in the kind (the weight fitted on the rows' own counts).
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 4
What Mindrift advertises it pays
How often does Mindrift state its pay, and how does it compare with other platforms for the same work?
| Roles | Stating an hourly rate | Its median per hour | Other platforms' median | Its median against theirs |
|---|---|---|---|---|
| All its open roles | 66.7% | $100 | $40 | +150.0% |
| Law, Medicine & Finance | 100.0% | $100 | $70 | +42.9% |
| Robotics & Physical-AI DataSmall sample | 0.0% | n/a | $10 | n/a |
| Agentic Tasks & RL EnvironmentsSmall sample | 100.0% | n/a | $65 | n/a |
Sample. 60 of the 90 roles Mindrift listed open on 31 Oct 2026 state an hourly rate, 19 distinct rates and 41 advertised maximums; other platforms: 3,505 of 4,291.
Pay is measured as on the job board: one sample per distinct advertised rate (the middle of a stated range; a role posted again at the same rate counts once), and an advertised maximum ("up to $X/hr") is a ceiling, not a rate, so it is counted and left out. Typical pay is the 25th percentile of what roles pay (the middle of each advertised range) to the 90th-percentile advertised ceiling; it and the median are rounded to $5 ($1 under $20), as on every list.
A median needs at least 3 distinct advertised rates, as on the job board. A comparison within a kind of work needs 3 distinct rates on each side. The row compared within the same kinds of work averages the kinds where both sides have enough, weighted by the platform's distinct rates in each, so a platform listing better-paid work is not read as paying more for the same work.
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 5
What Mindrift advertises it pays, week by week
Is Mindrift advertising higher or lower pay than it did?
| Week | Median per hour | Typical pay |
|---|---|---|
| 1 to 4 OctPartial week | n/a | n/a |
| 5 to 11 Oct | $60 | $50 to $80 |
| 12 to 18 Oct | $100 | $100 |
| 19 to 25 Oct | $100 | $60 to $100 |
| 26 to 31 OctPartial week | $100 | $60 to $100 |
Sample. Roles Mindrift listed open at the end of each ISO week (Monday to Sunday, UTC) that state an hourly rate, 1 Oct 2026 to 31 Oct 2026.
Pay is measured as on the job board: one sample per distinct advertised rate (the middle of a stated range; a role posted again at the same rate counts once), and an advertised maximum ("up to $X/hr") is a ceiling, not a rate, so it is counted and left out. Typical pay is the 25th percentile of what roles pay (the middle of each advertised range) to the 90th-percentile advertised ceiling; it and the median are rounded to $5 ($1 under $20), as on every list. Each week is measured over the roles open at its end.
A week needs at least 3 distinct advertised rates for a median.
A change can come from roles listed or removed at other rates as well as from a rate changed on a role: it describes what the board advertised, not what any one worker was paid.
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 6
Who may apply to roles at Mindrift
Are the roles Mindrift lists open to applicants anywhere, or limited to some countries?
| Location rule | Its open roles | Share of its roles | Share on other platforms |
|---|---|---|---|
| Open worldwide | 0 | 0.0% | 18.9% |
| Worldwide except listed countries | 0 | 0.0% | 18.5% |
| Only listed countries | 90 | 100.0% | 62.2% |
| No rule stated | 0 | 0.0% | 0.4% |
Sample. 90 roles Mindrift listed open on 31 Oct 2026, against the 4,291 open on other platforms.
The rule is read from each posting's location text, then, where that names no country, from what the posting's own description states, and where both are silent from what the platform states about where it hires. A rule says where applications are accepted; tax, payment and work-permit rules can still apply.
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 7
Countries Mindrift names in its limited roles
When Mindrift limits a role to some countries, which countries does it name?
| Country | Limited roles naming it | Share of its limited roles |
|---|---|---|
| United States | 49 | 54.4% |
| Jordan | 10 | 11.1% |
| Argentina | 7 | 7.8% |
| Bahrain | 7 | 7.8% |
| Brazil | 7 | 7.8% |
| Canada | 7 | 7.8% |
| Chile | 7 | 7.8% |
| Colombia | 7 | 7.8% |
| Czechia | 7 | 7.8% |
| Egypt | 7 | 7.8% |
| Hungary | 7 | 7.8% |
| India | 7 | 7.8% |
| Indonesia | 7 | 7.8% |
| Israel | 7 | 7.8% |
| Jamaica | 7 | 7.8% |
Sample. 90 of the 90 roles Mindrift listed open on 31 Oct 2026 accept applicants only from listed countries; 51 of them name a single country. The 15 countries named most often, of 62.
A role naming several countries counts for each. Open only there: roles whose list names this country alone.
A row is marked as a small sample where the platform's limited roles do not outweigh the average of the platforms for that country (the weight fitted on every platform's own count).
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 8
Listings posted and removed by Mindrift
How many roles did Mindrift post and take down?
| Measure | Roles |
|---|---|
| Open at the start of 1 Oct 2026 | 0 |
| New listings | 232 |
| Removed listings | 146 |
| Open at the end of 31 Oct 2026 | 90 |
Sample. Listings by Mindrift from 1 Oct 2026 to 31 Oct 2026.
New listings: roles first seen in the window, outside the platform's first week in this dataset and outside the week of a change in how this site reads its board. Removed: roles that left the feed during the window, other than with such a change. A role can be both.
On 14 Oct 2026 our reading of Mindrift changed: this site began reading its whole Workable board (950 rows, 399 distinct jobs) instead of its first 300 rows, its new postings judged over the next reads. From then to the end of that week (18 Oct 2026) the roles that came into view are counted as already listed, not as new listings and the roles that left view are not counted as removals, and its open roles are compared like for like across the change.
On 22 Oct 2026 our reading of Mindrift changed: this site began reading each posting's own text in its relevance check, and left out its three-day paid language studies in New York, which are not AI training work. From then to the end of that week (25 Oct 2026) the roles that left view are not counted as removals, and its open roles are compared like for like across the change.
The other platforms' count at the start leaves out platforms first tracked during the window.
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 9
How long roles at Mindrift stayed up
How long had the roles Mindrift removed been listed?
| Days listed | Removed roles | Share |
|---|---|---|
| Under 7 days | 31 | 21.2% |
| 7 to 13 days | 114 | 78.1% |
| 14 to 29 days | 1 | 0.7% |
| 30 to 59 days | 0 | 0.0% |
| 60 days or more | 0 | 0.0% |
Sample. 146 roles Mindrift removed from 1 Oct 2026 to 31 Oct 2026 with a known listing date; median 7.3 days (other platforms: 36.5 days over 1,752 roles).
Days listed: from the listing date the job board states (or the first day the role appeared in the feed) to the last day it was seen. Roles that arrived with the back catalogue and state no date are left out.
A median needs at least 3 removed roles. A removed role may have been filled, paused or closed: platforms do not say which.
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 10
What Mindrift calls the work
Which words does Mindrift use in its job titles: trainer, tutor, expert, rater, annotator?
| Title word | Its open roles | Share of its roles | Share on other platforms |
|---|---|---|---|
| Trainer | 42 | 46.7% | 18.1% |
| Engineer or developer | 41 | 45.6% | 7.3% |
| Evaluator or reviewer | 9 | 10.0% | 8.3% |
| Contributor or participant | 4 | 4.4% | 4.1% |
| Tutor | 1 | 1.1% | 0.4% |
| None of these words | 32 | 35.6% | 16.1% |
Sample. 90 roles Mindrift listed open on 31 Oct 2026, by the words in their titles (the same vocabulary as the job titles report).
A title can use several words ('Expert AI Trainer') and counts for each, so shares add to more than 100%. Whole words, case-insensitive.
A median needs at least 3 distinct advertised rates, as on the job board. Pay is measured as on the job board: one sample per distinct advertised rate (the middle of a stated range; a role posted again at the same rate counts once), and an advertised maximum ("up to $X/hr") is a ceiling, not a rate, so it is counted and left out. Typical pay is the 25th percentile of what roles pay (the middle of each advertised range) to the 90th-percentile advertised ceiling; it and the median are rounded to $5 ($1 under $20), as on every list.
A word is marked as a small sample where Mindrift's open roles do not outweigh the average of the platforms for that word (the weight fitted on every platform's own count).
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 11
Expert fields Mindrift hires for
Which professions and disciplines does Mindrift name in its job titles?
| Field | Its open roles | Share of its roles | Its share of all roles naming the field |
|---|---|---|---|
| Engineering (non-software) | 41 | 45.6% | 18.6% |
| Medicine and nursing | 19 | 21.1% | 8.3% |
Sample. 60 of the 90 roles Mindrift listed open on 31 Oct 2026 name a field in the title (2 fields).
Read from titles with the vocabulary of the expert fields report (whole words); a title naming two counts for both, and a role that needs a field without naming it is not counted.
A median needs at least 3 distinct advertised rates, as on the job board. Pay is measured as on the job board: one sample per distinct advertised rate (the middle of a stated range; a role posted again at the same rate counts once), and an advertised maximum ("up to $X/hr") is a ceiling, not a rate, so it is counted and left out. Typical pay is the 25th percentile of what roles pay (the middle of each advertised range) to the 90th-percentile advertised ceiling; it and the median are rounded to $5 ($1 under $20), as on every list.
Rows with fewer than 52.6 open roles naming the field are marked as a small sample and not ranked: below that the average of the rows weighs more than a row's own open roles naming the field (the weight fitted on the rows' own counts).
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 12
What roles at Mindrift ask for
Do the roles Mindrift lists ask for degrees, credentials, experience, set hours or tests?
| The description | Its roles | Share of its roles read | Other platforms with descriptions |
|---|---|---|---|
| Mentions a PhD or doctorate | 41 | 45.6% | 16.7% |
| Mentions a PhD or master's degree | 41 | 45.6% | 24.7% |
| Mentions a degree | 41 | 45.6% | 38.5% |
| Mentions a professional credential | 40 | 44.4% | 16.0% |
| States years of experience | 0 | 0.0% | 26.5% |
| States weekly hours | 60 | 66.7% | 23.6% |
| Mentions an assessment, test or trial task | 0 | 0.0% | 20.4% |
| Mentions an interview of any kind | 0 | 0.0% | 10.1% |
| Mentions an AI-led or recorded interview | 0 | 0.0% | 5.1% |
| Mentions a background or identity check | 0 | 0.0% | 1.4% |
| Mentions contractor or freelance terms | 41 | 45.6% | 52.6% |
Sample. The full descriptions of 90 of the 90 roles Mindrift listed open on 31 Oct 2026, against 3,910 roles of the other platforms whose boards publish them (AfterQuery, Aligned Labs, Anthropic, Anyone AI, Appen, Babel Audio, Braintrust, Claru, Cohere, DataAnnotation, DATAmundi, Edison Scientific, Elsevier Health, EvolutionIQ, Fleet AI, Handshake AI, Headway, iMerit, Innodata, Invisible, Inworld AI, JazzX AI, Kastle, Lifted (Upwork), Limbic, Mercor, micro1, Mistral AI, OneForma, OpenTrain AI, Outlier (Scale AI), Pasiflora AI, Perle Labs, Pila8, PointClickCare, Prolific, RealTrainAI, RemoExperts, Remotasks, Rise Data Labs, RWS TrainAI Community, Sama, Scale AI Human Frontier Collective, Shaip (Collaborate by Shaip), Sigma AI, Snorkel AI Expert Community, Sovrano AI, SuperAnnotate (SME), Surge AI, TELUS Digital AI, Terac, TransPerfect DataForce, Trellis, Turing, Vals AI, Vetto, Volga Partners, Weekday, Welo Data and xAI (SpaceXAI)). Median weekly hours stated: 10.
Read with the documented patterns of the requirements report (whole words, case-insensitive). A mention can be a requirement or a preference ('a PhD is a plus'): both count.
Medians of years and hours need at least 3 roles stating them (the one minimum of every median); for hours the lower figure of a range counts.
Rows with fewer than 1.7 described roles are marked as a small sample and not ranked: below that the average of the rows weighs more than a row's own described roles (the weight fitted on the rows' own counts).
Source. AITraining.jobs Research, Mindrift Hiring Data, October 2026 edition.
Figure 13
What an application to Mindrift asks for
What does an application to Mindrift ask for, and what must be filled in?
| Item | Its forms asking | Its forms requiring it |
|---|---|---|
| Cover letter | 100.0% | 0.0% |
| Location or country | 100.0% | 100.0% |
| Phone number | 100.0% | 100.0% |
| Resume or CV | 100.0% | 100.0% |
| Current employer | 0.0% | 0.0% |
| Other file uploads | 0.0% | 0.0% |
| GitHub profile | 0.0% | 0.0% |
| LinkedIn profile | 0.0% | 0.0% |
| Portfolio or website | 0.0% | 0.0% |
Sample. The application forms of 90 of the 90 roles Mindrift listed open on 31 Oct 2026, read from its applicant-tracking system (Workable); a median of 7 fields and 0 questions of its own (other platforms' 1,592 forms: 10 fields).
Required: the form will not submit without it. Name and email, asked by every form, are not listed. Some platforms apply in steps (a short form, then an assessment); only the first form is read.
No row with any forms is a small sample: the rows differ so much that each reads on its own from its first.
Source. AITraining.jobs Research, Mindrift Hiring Data, October 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:
11,482 roles tracked, each with its history
Every role on 62 job boards, read since 13 Jul 2026 (every hour since 18 Oct 2026): its pay and place, and the first and last day it was open.
58,632 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. Mindrift's figures come from its own public listings only; full descriptions and application forms are read where its job board publishes them. 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.
- Other platforms
- Every platform this site reads other than Mindrift, over the same window and pooled: counts are summed and medians are taken over all their rates together. Mindrift is compared with the rest of the market, not with a market that includes it.
- Rank
- Its place among the platforms with open roles on the last day of the window, highest first (longest first for days listed). A share is ranked where the platform's open roles read on their own among the platforms (they reach the weight of the platforms' average share, fitted on their own counts), a pay median from 3 distinct advertised rates and days listed from 3 removed roles, so a platform below one is not ranked on that measure.
- Open role
- A role listed by a given day and still in its platform's feed that day. Week by week, open roles are counted at the end of each week (Sunday, UTC).
- Back catalogue
- The roles already listed in the week a platform was first tracked. That week is marked in the weekly series and its roles are not counted as new listings; weeks before it are left out.
- New listing
- A role that first appeared during the window on a platform already tracked before that week, other than with a change in how this site reads its board (below).
- Removed listing
- A role that left its platform's feed during the window, other than with a change in how this site reads its board (below). When most removals or new listings fall on one day, the figure says so: one large change to a board moves the window's flow.
- Change in our reading
- 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. For Mindrift: on 14 Oct 2026 this site began reading its whole Workable board (950 rows, 399 distinct jobs) instead of its first 300 rows, its new postings judged over the next reads; on 22 Oct 2026 this site began reading each posting's own text in its relevance check, and left out its three-day paid language studies in New York, which are not AI training work.
- Days listed
- From a removed role's listing date to the last day it was seen in its feed. A median needs at least 3 removed roles with a known listing date.
- Advertised pay
- Measured exactly as the job board measures any list of roles, so a median here is the median Mindrift's pay page states for the same roles: one sample per distinct advertised rate (the middle of a stated range; a role posted again at the same rate counts once), and an advertised maximum ("up to $X/hr") is a ceiling, not a rate, so it is counted and left out. Typical pay is the 25th percentile of what roles pay (the middle of each advertised range) to the 90th-percentile advertised ceiling; it and the median are rounded to $5 ($1 under $20) and need at least 3 distinct rates.
- Compared within the same kinds of work
- For each kind of work where both the platform and the other platforms state at least 3 distinct hourly rates, its median over theirs, minus one; averaged across those kinds, weighted by the platform's distinct rates, so a platform listing better-paid work is not read as paying more for the same work. It needs two such kinds.
- Location rule
- Read from each posting's location text and, where that names no country, from what the posting's own description states: open worldwide, worldwide except listed countries, only listed countries, or no rule stated.
- 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.
- Title word, language, field
- Read from titles as whole words with the vocabularies of the job titles, languages and expert fields reports; a title can count for several.
- Mention
- A description mentions a degree, credential, step or term when it contains one of the patterns published in the requirements report. A mention may be a requirement or a preference.
- Application form
- The fields an applicant fills in on the platform's own application page, read from its applicant-tracking system. Required: the form will not submit without the field.
- Window
- Live: the last 28 complete days, in UTC, with the weekly series over the last twelve complete ISO weeks from the week of 3 Aug 2026 at the earliest. Edition: one calendar month, in UTC, its weeks cut at the month's edges.
What this data cannot tell you
- Platforms that recruit without a public job board are not counted, and a platform may staff projects it never lists: its listings are what it advertises, not everything it hires for.
- A listing is not a hire. Some platforms keep evergreen roles open for months, and some post one listing per language or country.
- Listings only: how many people apply to a platform is never published here.
- Pay is what postings advertise, not what workers are paid after bonuses, tiers or rejected tasks.
- Full descriptions and application forms are published by some job boards only, so those figures appear only for the platforms whose boards publish them, and describe the roles open on the day they were read.
- A removed listing may have been filled, paused or closed: platforms do not say which.
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 28 complete days and changes every day. 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.
- Live dataLast 28 complete days, recomputed daily
- October 2026 edition (this page)Freezes on first reading
- September 2026 editionFrozen 10 Oct 2026
- August 2026 editionFrozen 10 Oct 2026
Cite and reuse
Free to quote, chart and republish under CC 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). Mindrift Hiring Data, October 2026 edition. https://aitraining.jobs/research/mindrift-hiring-data/2026-10
BibTeX
@misc{aitrainingjobs_mindrift_hiring_data_2026_10,
author = {{AITraining.jobs}},
title = {Mindrift Hiring Data, October 2026 edition},
year = {2026},
howpublished = {\url{https://aitraining.jobs/research/mindrift-hiring-data/2026-10}},
}More research
The same figures for every other platform:
- OpenTrain AI hiring data
- Invisible hiring data
- micro1 hiring data
- Welo Data hiring data
- AfterQuery hiring data
- Mercor hiring data
- Sigma AI hiring data
- Braintrust hiring data
- Alignerr hiring data
- TransPerfect DataForce hiring data
- Appen hiring data
- xAI (SpaceXAI) hiring data
- Terac hiring data
- SuperAnnotate (SME) hiring data
- Turing hiring data
- Handshake AI hiring data
- Anyone AI hiring data
- Perle Labs hiring data
- Claru hiring data
- Babel Audio hiring data
- RemoExperts hiring data
- Which AI Labs Are Hiring for Human DataWhich AI labs are hiring for the teams that buy, run and evaluate human data, and for what roles?Read the report
- Who Is Buying AI Training DataWhich AI labs and data vendors have people whose job is to source, license or acquire training data right now?Read the report
- Where Human-Data Vendors Are BuildingHow fast are the human-data vendors hiring their own staff, and for which functions?Read the report
- What AI Labs Pay for Human DataWhat do AI labs spend on human data, what do vendors bill, what do the experts receive, and what do RL environments and tasks cost?Read the report
- What AI Labs Are BuyingWhich kinds of training data are in rising demand, and which are cooling?Read the report
- AI Training Data Licensing DealsWho has licensed content and data to which AI companies, on what terms and for how much?Read the report
- The State of AI Training PayWhat does AI training work pay, by kind of work and by platform?Read the report