Last Updated on October 2, 2026 by Justin Bryant
AI-training platforms can advertise impressive pay and flexible remote work, but that does not answer the question most applicants actually have: which platforms appear to have opportunities available right now?
I checked seven major AI-training and AI-data platforms. Together, they showed 1,733 visible public listings on October 2, 2026. Meridial had the largest catalog with 833. Mercor had the strongest combination of meaningful scale, recent momentum, and overall platform quality. Mindrift had the highest freshness ratio in my September 30 tracking snapshot.
Those numbers need an important warning. A public listing proves that an application route exists. It does not prove that you will be accepted, receive a project, find tasks in your private dashboard, or earn steady income. Geographic duplicates, repeated templates, and evergreen talent pipelines can also make a platform look busier than it really is.
AI-training platforms with the most visible listings
| Rank | Platform | Visible listings | Newly detected in 14 days* | SMS score | Best interpretation |
|---|---|---|---|---|---|
| 1 | Meridial | 833 | 1 | 59% — D | Largest catalog, but many geographic variants and little recent movement |
| 2 | Mercor Experts | 328 | 155 | 74% — A | Best overall option for qualified specialists |
| 3 | Handshake AI | 209 | 7 | 54% — E | Large and transparent, but restricted by eligibility and expertise |
| 4 | DataForce | 148 | Not consistently exposed | 65% — B | Accessible mixed projects, often temporary or country-specific |
| 5 | Rise Data Labs | 116 | 0 | 49% — F | Broad expert catalog with weak recent movement |
| 6 | Mindrift | 92 | 87 | 62% — C* | Strongest freshness ratio in the September 30 snapshot |
| 7 | Outlier | 7 | 0 | 58% — D | Small public board; private contributor task volume remains unknown |
Visible listing totals were rechecked on October 2, 2026. The 14-day figures are from the complete September 30 comparison snapshot and can change daily. Mindrift's 62% is a provisional score explained below.
The count changed quickly. These same platforms showed 1,777 visible listings in the September 30 snapshot. By October 2, Mercor had fallen from 375 to 328 listings, Outlier moved from eight to seven, Handshake rose from 207 to 209, DataForce rose from 147 to 148, and Meridial rose from 832 to 833. That two-day movement is exactly why I would treat every count as a dated snapshot rather than a permanent ranking.
How I compared the platforms
I used two different kinds of evidence because they answer different questions:
- Visible inventory: contributor-facing roles or opportunities that an outside applicant could see.
- Recent momentum: roles newly detected during the prior 14 days when the independent tracker exposed a comparable figure.
I also included each platform's existing Self-Made Success score. That score considers barrier to entry, work consistency, pay transparency, work difficulty, flexibility, and reputation. Reputation supplies half of the weighted result. A higher work-difficulty score means the work is easier or more manageable for a qualified contributor.
The overall score is useful context, but it is not a measurement of current listings. That is why Meridial can lead the public inventory ranking while receiving a 59% overall score, and why Mercor can have fewer listings but earn the strongest overall rating at 74%.
7. Outlier: 7 visible opportunities
Best for: Applicants whose skills and country match one of the currently visible domains.
Biggest catch: Outlier is a recognizable platform with a large contributor network, but its public opportunity page is currently much smaller than the other boards in this comparison.
Outlier's official opportunity page showed seven visible roles in my October 2 check. The prior September 30 snapshot showed eight roles and no newly detected listing in the preceding 14 days. The visible mix included generalist, coding, language, voice-evaluation, and specialist opportunities.
This does not prove that existing Outlier contributors have no tasks. Public recruiting inventory and private dashboard supply are different things. I would interpret the count as a weak signal for a new outside applicant—not a platform-wide declaration that no work exists.
Existing SMS score: 58% — D tier.
- Barrier to entry: 2/5
- Work consistency: 2/5
- Pay transparency: 4/5
- Work difficulty: 3/5
- Flexibility: 3/5
- Reputation: 3/5
The score reflects meaningful screening, unpredictable project access, and recurring availability and support complaints. Outlier generally does better on visible rate information than it does on dependable task supply.
Read my full Outlier AI review
6. Mindrift: 92 visible roles
Best for: Remote applicants who match a current specialty or want to test Mindrift's newer all-access route.
Biggest catch: A burst of new listings does not establish dependable work after onboarding.
Mindrift's application route showed 92 tracked roles. In the September 30 freshness snapshot, 87 of those 92 had been newly detected in the preceding 14 days, and all 92 were labeled remote. That was the strongest freshness ratio in the comparison.
The roles covered areas such as coding, design, writing, languages, creative work, and reasoning across multiple countries. Mindrift also introduced an all-access route for lighter tasks in July 2026. That route is different from qualifying for a specialist project, and it should not be combined with expert-level advertised rates as if every user receives the same work.
Mindrift itself cautions that AI-training work should not be treated as stable full-time employment. Advertised rates are commonly framed as “up to” amounts rather than guaranteed effective earnings.
Existing SMS score: provisional 62% — C tier.
- Barrier to entry: 4/5
- Work consistency: 2/5
- Pay transparency: 3/5
- Work difficulty: 3.5/5
- Flexibility: 3.5/5
- Reputation: 3/5
The provisional score reflects the easier all-access route. It does not raise work consistency above 2 out of 5 because easier entry and fresh applications do not guarantee an ongoing task queue.
Read my complete Mindrift review
5. Rise Data Labs: 116 visible roles
Best for: Experienced professionals willing to search for a close match by specialty, location, credential, and project requirements.
Biggest catch: A large, mostly remote catalog can still represent evergreen recruiting pipelines rather than 116 separate active client projects.
Rise Data Labs' official jobs page exposed 116 unique role links. The independent September 30 snapshot also counted 116, with 115 labeled remote and no newly detected role during the preceding 14 days.
The board covered finance, science, coding, creative work, media, professional services, government-related expertise, and other specialist domains. Some opportunities were global. Others required a specific country, professional background, software tool, degree, or number of years of experience.
Existing SMS score: 49% — F tier.
- Barrier to entry: 2/5
- Work consistency: 2/5
- Pay transparency: 3.5/5
- Work difficulty: 2/5
- Flexibility: 2.5/5
- Reputation: 2.5/5
The low score is not a claim that Rise is fake. It reflects demanding qualifications, uncertain continuity, intensive expert work, short project windows that may request 20 to 40 hours per week, and a limited independent contributor record.
4. DataForce by TransPerfect: 148 tracked roles
Best for: Applicants who are open to speech, photo, video, language, research, annotation, and data-collection projects—not only evaluating chatbot answers.
Biggest catch: Country variants and one-time collection projects inflate the raw inventory, and some projects are onsite rather than remote.
The official DataForce Community covers data sourcing, data annotation, user studies, and job opportunities. The independent tracker showed 148 open listings on October 2, up from 147 in the September 30 snapshot.
This is the most varied project mix in the comparison. Many opportunities have straightforward participation requirements, but eligibility can depend on your country, language, device, age, demographic profile, or ability to attend an onsite session. Projects can also close after the client collects enough approved submissions.
Existing SMS score: 65% — B tier.
- Barrier to entry: 4/5
- Work consistency: 2/5
- Pay transparency: 3.5/5
- Work difficulty: 3.5/5
- Flexibility: 4.5/5
- Reputation: 3/5
DataForce earns its B-tier score from accessibility and flexibility—not from dependable weekly assignments. Many projects are temporary, region-specific, and paid only after submitted work passes quality review.
3. Handshake AI: 209 visible roles
Best for: Eligible US-based students, graduates, academics, and professionals with a matching specialty.
Biggest catch: Location, work authorization, educational background, expertise, and project assessments greatly reduce how many of the raw listings will be relevant to one applicant.
The Handshake AI opportunity board had 209 tracked roles on October 2, up from 207 in the September 30 snapshot. All of the tracked roles displayed pay information, and the visible categories ranged from generalist projects to engineering, medicine, science, logistics, design, and business specialties.
Handshake's help documentation says availability can change quickly and that projects may impose weekly hour caps. Current rules also make geography essential: US applicants need eligible work authorization and tax identification, while non-US access may depend on an invitation.
Existing SMS score: 54% — E tier.
- Barrier to entry: 2/5
- Work consistency: 2.5/5
- Pay transparency: 3/5
- Work difficulty: 3/5
- Flexibility: 4/5
- Reputation: 2.5/5
The board is large and unusually transparent about rates, but that does not erase its eligibility limits, matching risk, or mixed contributor-specific reputation.
Read my full Handshake AI review
2. Mercor Experts: 328 visible roles
Best for: Experienced professionals with a clear specialty who are comfortable completing assessments and an AI interview.
Biggest catch: Strong volume does not make Mercor beginner-friendly. Many roles are narrow, credentialed, or restricted by location.
Mercor Experts had 328 tracked roles on October 2. That was down from 375 in the September 30 snapshot, when 155 roles had been newly detected during the preceding 14 days and 341 of the 375 were labeled remote.
The network included law, medicine, finance, engineering, operations, research, languages, and AI evaluation. Mercor's talent documentation explains the assessment process, AI interview, supported countries, contracts, time tracking, and weekly payments. Some opportunities pay hourly, while others use per-task or deliverable terms, so there is no single platform-wide rate.
Existing SMS score: 74% — A tier.
- Barrier to entry: 2/5
- Work consistency: 3/5
- Pay transparency: 4/5
- Work difficulty: 3/5
- Flexibility: 2.5/5
- Reputation: 4.5/5
Mercor has the highest existing SMS score in this group and the best combination of scale and recent momentum for specialists. Its 2-out-of-5 entry score is the reason I would not call it the easiest first option for a beginner.
Read my complete Mercor review
1. Meridial: 833 visible roles
Best for: Applicants who can filter a very large catalog to an exact country, language, or professional-domain match.
Biggest catch: The largest number is also the strongest example of why listings should not be confused with distinct live assignments.
Meridial's public project board led the comparison with 833 tracked listings on October 2, up from 832 on September 30. The board covered law, STEM, finance, coding, languages, safety, medicine, creative fields, and generalist work.
Only one listing had been newly detected during the 14-day window used in the September 30 comparison. Many opportunities were also country or specialty variants: 494 of the 832 listings in that snapshot were categorized as region-specific. Pay appeared on nearly all of them, but broad ranges repeated across many related listings.
Existing SMS score: 59% — D tier.
- Barrier to entry: 3/5
- Work consistency: 2/5
- Pay transparency: 3/5
- Work difficulty: 4/5
- Flexibility: 2.5/5
- Reputation: 3/5
Meridial wins the inventory count. It does not win the overall score because joining a large expert network can still produce long matching gaps, and individual projects may carry schedule or availability expectations.
Read my complete Invisible Technologies and Meridial review
Which AI-training platform should you try first?
- Experienced professional: Start with Mercor, then filter Meridial and Handshake for a close domain and location match.
- Applicant prioritizing newly posted remote roles: Check Mindrift first, while remembering that its work-consistency score remains 2 out of 5.
- Beginner open to one-time data projects: DataForce may be more accessible than the expert networks, especially when your language, location, device, or demographic profile matches a project.
- US student, graduate, or subject expert: Handshake may be worth checking, but confirm work authorization and project eligibility before investing time in assessments.
- Existing Outlier contributor: Check your private dashboard rather than assuming the seven-role public page represents your current task access.
I would not apply to all seven blindly. Start with two or three platforms that match your background, country, preferred project type, and schedule. A smaller board with a precise match can be more valuable than hundreds of listings you cannot qualify for.
Frequently asked questions
How many AI-training openings did these platforms have?
The seven platforms showed 1,733 visible public listings when rechecked on October 2, 2026. The comparable September 30 snapshot contained 1,777. The difference shows how quickly public inventories can change.
Which AI-training platform had the most open listings?
Meridial had the largest public catalog with 833 tracked listings on October 2. Many were country or specialty variants, however, and the September 30 audit detected only one new listing during the preceding 14 days.
Which platform appeared to be adding jobs fastest?
Mindrift had the highest freshness ratio in the complete September 30 snapshot: 87 of 92 tracked roles had been newly detected during the prior 14 days. Mercor added more roles in absolute terms—155—but had a much larger catalog.
Which platform ranked best overall?
Mercor received the highest existing Self-Made Success score at 74%, or A tier. It combined strong reputation, pay transparency, and potential project duration with a large board and strong recent momentum. Its low entry score means this recommendation is mainly for qualified specialists.
Do public listings mean work is immediately available?
No. A listing can represent a recruiting pipeline, repeated geography variant, or project that still requires screening and client matching. Acceptance also does not guarantee tasks, hours, or income.
Are these employee jobs?
Many opportunities are independent contractor, project, fellowship, task, or expert-network arrangements rather than ordinary full-time employment. Read the specific listing and contract carefully before applying.
Why is a platform with hundreds of listings scored below one with fewer?
The SMS score measures the broader applicant experience, including entry barriers, consistency, pay visibility, difficulty, flexibility, and reputation. Listing volume answers a narrower question. A large catalog can coexist with difficult screening, weak matching, or inconsistent assignments.
My final verdict
Meridial has the largest catalog, Mindrift had the strongest freshness ratio, and Mercor is the best overall platform in this seven-company comparison. Those are three different wins because they answer three different questions.
If I were an experienced professional, I would begin with Mercor, then check Meridial for an exact specialty and geography match. If I cared most about newly appearing remote roles, I would check Mindrift early. If I wanted lower-barrier, short data projects, I would investigate DataForce.
The safest conclusion is not that 1,733 jobs are available to everyone. It is that these seven boards displayed roughly 1,700 public application opportunities on the evidence date—and that only a fraction will be relevant to any one applicant.
You can also compare a broader set of platforms in my 14-platform AI-training ranking or use the Self-Made Success Side Hustle Finder to filter opportunities by entry difficulty, consistency, pay transparency, flexibility, and reputation.