Welo Data
Welocalize's AI-data arm: rater, search-evaluation and language roles locked to a country and language.
Last reviewed 2026-10-14
Typical pay
Generative AI Analyst $3–28/hr by country; US raters about $20/hr
Who can join
Postings are locked to a country and language; ID and tax checks; no VPNs
How you get in
Account, language assessment, unpaid task or client exam, then agreement
Payouts
Not published; confirmed at onboarding
Watch out
Unpaid testing, 'project is full' reports, very low rates in some countries, recruiter scams
What it is
- Official
Welo Data is the AI training-data brand of Welocalize, a localization company founded in 1997. Welocalize launched the Welo Data brand on May 21, 2024, led by SVP and General Manager Siobhan Hanna. Services named at launch: annotation and labeling, prompt engineering, RLHF, red teaming, data collection and search/ads relevance work.
silicon.co.uk· checked 2026-10-14 - Official
In April 2026 the group adopted Welo Global as its parent brand over Welocalize, Welo Data, Park IP, Adapt and the new Welo Life Sciences. The release names Paul Carr as CEO and claims more than 2,000 clients, 300+ languages and over 500,000 linguists and domain experts.
ciodive.com· checked 2026-10-14 - Official
Welo Data presents itself to clients as a managed service, "not a self-serve marketplace", with "500k+" curated experts, 7 ISO certifications and SOC 2.
welodata.ai· checked 2026-10-14 - Official
Contributor-facing copy claims "100+ Countries" and 500,000+ contributors. These are company figures.
welodata.ai· checked 2026-10-14
Coding work
- Official
Little coding work for remote contributors. A single read of Welo Global's Lever board on 2026-10-14 returned 537 postings. No remote, freelance coding or code-evaluation gig was listed. The coding-adjacent roles found are below.
api.lever.co· checked 2026-10-14 - Official
Senior Software Engineer, Robot Data Collection ("Project Hydra"): $105/hour, but it is a full-time W-2 role, 100% onsite in Santa Clara, CA, Monday to Friday, on a 12-month contract. It asks for 3+ years building software for real robot systems, strong Python and Linux, and ROS 2 or similar. This is an engineering job running a data-collection fleet, not gig work.
jobs.lever.co· checked 2026-10-14 - Official
Senior Prompt Engineer ("Project Lion"): remote, part-time, freelance, posted for France, Germany, Italy, Japan, Netherlands, Portugal and Spain. Work is migrating rating templates to LLM autoraters, writing and testing prompts, and measuring precision/recall against gold data. It asks for a degree in CS, data science, computational linguistics or similar, 4+ years as a prompt engineer, SQL or similar, and native fluency in the local language. No pay is listed.
jobs.lever.co· checked 2026-10-14 - Official
No coding needed for most work: contributor postings say "No coding required, just your expertise".
jobs.lever.co· checked 2026-10-14
Show 13 moreShow less
- Official
Board composition: by team, the 537 postings were "AI Service General Application" 204, Data Annotation 109, Data Validation 65, Production 58, Translation & Localization 37, plus smaller teams. 163 postings carried a pay range. Postings covered 83 countries, led by the US (114), India (46) and Spain (46).
api.lever.co· checked 2026-10-14 "AI Trainers Network" talent-pool entries for about 100 languages: not an active opening; you join a database and are contacted with scope and pay when a project fits. Work types named: annotation, search/ads/chatbot evaluation, prompt creation and testing, and cultural review.
jobs.lever.co· checked 2026-10-14Generative AI Analyst (text, image, audio and video evaluation): from $3/hr (Vietnamese) to $28/hr (Dutch, Netherlands); French (France) is $18/hr, freelance, 4 weeks.
Voice and speech data ("Hydrus"): voice contributors $20–$65/hr and session directors $50–$80/hr depending on locale.
"Project Perseus" US roles: data quality analyst $32–$38/hr, data labeling associate $34/hr, speech and voice analyst $28/hr. The analyst listing is onsite (Bay Area, Seattle, NYC), full-time W-2, 40 hours a week, US work authorization required with no visa sponsorship.
jobs.lever.co· checked 2026-10-14Maps relevance raters: from $3.30/hr (English, India) and $4/hr (Indian languages) to $20.30/hr (US, Canada) and $29/hr (US visual design).
Translation validators: $3.50/hr (several Indian languages) to $14.30/hr (Spanish, Spain).
jobs.lever.co· checked 2026-10-14PhD specialists (Japanese, Korean) in social sciences, humanities, arts and linguistics: $50–$70/hr, posted as a pipeline for a project expected to start months later.
jobs.lever.co· checked 2026-10-14Brazilian PhD-level subject experts (law and ethics, history, politics, literature, philosophy, sports, animals) write expert questions and cited answers in Portuguese; no pay listed.
jobs.lever.co· checked 2026-10-14Customs and trade compliance subject expert (Mexico): $12/hr.
jobs.lever.co· checked 2026-10-14Robot manipulation video annotation (Philippines): $3.50–$4.03/hr.
jobs.lever.co· checked 2026-10-14Lowest rate seen: Dutch data annotator in Suriname at $2.25/hr.
jobs.lever.co· checked 2026-10-14- Official
Contract types vary by posting: freelance/self-employed (for example German Ads Quality Rater, "Remuneration: Variable depending on the complexity of the task", 5–20 hours a week) and W-2 part-time employee (US Search Quality Rater).
Eligibility
- Official
Country-locked postings: each listing names a country or region and a language (for example "Must be located in Canada"). There is no single published country list.
jobs.lever.co· checked 2026-10-14 - Official
US rater states: the US Search Quality Rater listing hires only in 19 named states (including Florida, Texas, Pennsylvania, Ohio and Virginia) and requires age 18+.
jobs.lever.co· checked 2026-10-14 - Official
Exclusivity rules: search raters must work only on the search rating program, must not have current or past ads-rating experience, and must be the only search rater in the household. An NDA and a client quality test are required.
jobs.lever.co· checked 2026-10-14 - Official
Identity checks: Welo Data says it runs "numerous identity verification checks" and tells applicants not to use IP obfuscation tools. Onboarding happens "once verified (ID & Tax only)".
Show 2 moreShow less
- Official
Equipment: a computer, stable internet and a modern browser; some projects need extra software or skills.
welodata.ai· checked 2026-10-14 - Official
Visa: US onsite roles require US work authorization with no visa sponsorship.
jobs.lever.co· checked 2026-10-14
How to get in (format only)
- Official
Steps: (1) apply for a role; (2) create a Welo Works account from the email you receive; (3) take a language and cultural assessment; (4) optional study materials; (5) pass a task-related exam; (6) after approval, submit payment details and sign the agreement.
welodata.ai· checked 2026-10-14 - Official
Rater projects: screening questions, candidate portal registration, a skill and language evaluation, "Know-How" guides, a client entrance exam, then onboarding after ID and tax verification. The process is described as "self-directed and automated".
- Official
Scam warning: legitimate messages come only from @welocalize.com, @weloglobal.com, @hire.lever.co or @welodata.ai, and Welo Data says it never charges fees or asks for equipment purchases.
welodata.ai· checked 2026-10-14 - Worker reports
Trustpilot reviewers (Jul to Oct 2026) describe hours of unpaid study and exams, failed results with no score shown, and being locked out after onboarding.
trustpilot.com· checked 2026-10-14
Pay and payments
- Official
Official position: pay is per task or project and varies by task type, complexity and volume. Payments go through "secure, verified online payment platforms", and the exact method and frequency are confirmed during project onboarding. Contributors see task history, rates and earnings in a dashboard.
welodata.ai· checked 2026-10-14 - Official
W-2 roles: the US Search Quality Rater role pays "every 2 weeks" and includes paid sick time and, for eligible workers, medical, dental, vision and 401(k).
jobs.lever.co· checked 2026-10-14 - Official
Rates by country: the same role pays very different rates by location; Generative AI Analyst ranges from $3/hr to $28/hr across locales (see section 2).
api.lever.co· checked 2026-10-14 - Worker reports
Trustpilot reviewers report a completed project never paid (Jun 2026) and job posts that hid word count and rate until after a claim (Aug 2026).
trustpilot.com· checked 2026-10-14
Work availability
- Official
Official position: projects last from a few days to several weeks; talent-pool sign-ups are "not an active job opening". Rater listings say "This work is based on project needs. Weekly hours may vary."
- Official
Hours bands: search raters 10–20 hours a week (Canada, freelance) or 10–29 hours a week (US, W-2); German ads raters 5–20 hours a week.
- Official
Pipeline postings: some listings recruit ahead of a project that has not started, such as the PhD specialist pipeline.
jobs.lever.co· checked 2026-10-14 - Worker reports
Trustpilot reviewers describe repeated "project is full" messages, jobs vanishing from the dashboard, and three months with no earnings after passing exams (Sep to Oct 2026).
trustpilot.com· checked 2026-10-14
Risks and controversies
- News
NLRB charge (2024): a charge against Welocalize was filed Nov 4, 2024 with NLRB Region 2 (New York), alleging discharge and retaliation for concerted activity. It was dismissed on Jun 12, 2025 and is closed. A charge is an allegation, not a finding.
nlrb.gov· checked 2026-10-14 - News
Layoff notice (2019): an Oregon WARN notice records Welocalize terminating 73 "Linguistic Reviewers" effective Aug 16, 2019.
ccwd.hecc.oregon.gov· checked 2026-10-14 - Worker reports
Deactivations: Trustpilot reviewers (Jul to Oct 2026) report deactivation after VPN/proxy warnings they say were false, a ban after submitting ID, and accounts deleted after contacting support, with support replies such as "We're unable to provide further details regarding this decision."
trustpilot.com· checked 2026-10-14 - Official
Impersonation scams: Welo Data warns about fake recruiters and lists its real sending domains (see section 4).
welodata.ai· checked 2026-10-14 - Official
Low rates in some countries: posted rates fall to $2.25–$4/hr in several locales (see section 2).
jobs.lever.co· checked 2026-10-14
Referral program
- Official
No contributor referral program found. Neither the contributor page nor the main site mentions one.
- Official
Staff (full-time corporate) postings list an "Employee Referral Program" as a benefit; this is for employees, not freelance contributors.
jobs.lever.co· checked 2026-10-14
Worker sentiment (reports, not verified)
- Worker reports
Ratings: Trustpilot (welocalize.com) shows 1.3/5 from 98 reviews, 95% 1-star; the profile is unclaimed and the company does not invite reviews, so it skews to complaints.
trustpilot.com· checked 2026-10-14 - Worker reports
Negative, top theme: opaque screening and deactivation with automated support replies.
trustpilot.com· checked 2026-10-14 - Worker reports
Negative: unpaid assessment time and little or no work after qualifying.
trustpilot.com· checked 2026-10-14 - Worker reports
Positive: employee reviews on Glassdoor praise supportive leadership and autonomy; little positive contributor feedback was found publicly.
glassdoor.ca· checked 2026-10-14
Quick answers
- How much does Welo Data pay?
- Generative AI Analyst $3–28/hr by country; US raters about $20/hr
- Who can join Welo Data?
- Postings are locked to a country and language; ID and tax checks; no VPNs
- How do you get into Welo Data?
- Account, language assessment, unpaid task or client exam, then agreement
- How and when does Welo Data pay out?
- Not published; confirmed at onboarding
- What should you watch out for at Welo Data?
- Unpaid testing, 'project is full' reports, very low rates in some countries, recruiter scams