Skip to content
14 roles liveSalary guide →

AI training

Get paid to train the world’s best AI

Frontier models depend on people who write, review, and grade their work. BridgeWide matches you to the platforms that pay for that work, across software, data, writing, STEM, and other specialist tracks. You apply once. The platform pays you.

  • Apply once
  • Reviewed in three business days
  • Paid by the platform
People working together at a tableApply to brief · 3 days

The directory

18 platforms, 5 regions

The work

Three kinds of tasks, and the tracks around them.

You apply once. The platform pays you.

01 / 03

Code evaluation and agentic tasks

Compare model-written solutions, find the bugs, and write why one answer is better. Design realistic multi-file tasks and grade how a coding agent works through them.

Pay
USD 8–150/hour
Roles
8
Regions
United States, Canada, Latin America, Europe
Where
Remote

Mission

Close the gap.

Frontier AI is built on expert human judgement, but people across the United States and the rest of the world never hear about the platforms that pay for it. BridgeWide exists to close that gap.

At a glance

One application, paid by the platform.

  • 00

    Platforms in the directory

  • Worldwide

    Remote, including the United States

  • $0

    Cost to candidates

  • 0

    Business days to review

What you can count on

A woman filling in a form on her laptop

For candidates01 / 04

What you can count on

One application.

A single form and CV cover every platform in the directory. Tell the desk your discipline, region, English level, and hours, and it matches from there.

Apply once

3 min

to fill in, CV attached

Promise 01

The desk aims to review it within three business days.

The path

Apply, match, brief, start.

The desk aims to review an application within three business days.
Apply once
  1. 01/ 04

    Apply once

    Send your experience, the kind of work you do, your region, your English level, and a CV. It takes about three minutes.

    Then: Get screened and matched

  2. 02/ 04

    Get screened and matched

    We read the profile and match it to platforms by discipline, seniority, language, and time zone.

    Then: Get briefed and introduced

  3. 03/ 04

    Get briefed and introduced

    You get a briefing on the assessment. Where BridgeWide has a direct line, we make the introduction. Your profile is shared only after you agree.

    Then: Start on the platform’s terms

  4. 04/ 04

    Start on the platform’s terms

    The platform pays you, in USD, on its own terms. Part-time is about ten hours a week, and full-time is available. BridgeWide does not take a cut.

    Paid by the platform

Roles

The range of the work.

From $8 to $180 an hour, in USD, paid by the platform. Each bar is one role's range.
  1. RLH

    Generalist writing and rating

    United States · RLHFRate and rewrite model answers. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 25–50/hour
  2. COD

    Coding and software evaluation

    United States · CodeEvaluate code and software tasks. The published top is $150 or more an hour. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 40–150/hour
  3. SPC

    Multilingual rating

    Named on the project · SpecialistRate answers across languages. The project names the country. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 25–40/hour
  4. SPC

    STEM and professional evaluation

    United States · SpecialistSTEM, finance, legal, and other professional evaluation. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 40–125/hour
  5. COD

    Software projects

    United States · CodeSoftware projects on Mercor's talent network. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 70–150/hour
  6. SPC

    Business expert projects

    United States · SpecialistBusiness expert projects. Task work can be priced separately by the platform. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 50–100/hour
  7. SPC

    Healthcare expert projects

    United States · SpecialistHealthcare expert projects for credentialed clinicians. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 50–180/hour
  8. SPC

    Spanish tutoring

    Latin America · SpecialistSpanish tutoring for model training. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 60–65/hour
  9. RLH

    General inbound

    United States · RLHFGeneral inbound evaluation. Outlier publishes a ceiling, not a floor. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 35/hour
  10. COD

    Coding inbound

    United States · CodeCoding inbound. Outlier publishes a ceiling, not a floor. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 30/hour
  11. RLH

    General inbound

    Canada · RLHFGeneral inbound evaluation for people in Canada. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 35/hour
  12. COD

    Coding inbound

    Canada · CodeCoding inbound for people in Canada. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 30/hour
  13. RLH

    General inbound

    Latin America · RLHFGeneral inbound evaluation for people in Mexico. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 35/hour
  14. COD

    Coding inbound

    Latin America · CodeCoding inbound for people in Mexico. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 30/hour
  15. RLH

    General inbound

    Europe · RLHFGeneral inbound evaluation for people in the United Kingdom. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 35/hour
  16. COD

    Coding inbound

    Europe · CodeCoding inbound for people in the United Kingdom. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 30/hour
  17. RLH

    English voice conversation

    Canada · RLHFEnglish voice conversation evaluation. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 15/hour
  18. RLH

    English voice conversation

    Europe · RLHFEnglish voice conversation evaluation. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 15/hour
  19. SPC

    Android evaluator, Japanese

    Asia · SpecialistAndroid evaluation in Japanese. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 31/hour
  20. SPC

    Android evaluator, Korean

    Asia · SpecialistAndroid evaluation in Korean. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 31/hour
  21. SPC

    Android evaluator, French

    Europe · SpecialistAndroid evaluation in French. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 25/hour
  22. SPC

    Android evaluator, Spanish

    Latin America · SpecialistAndroid evaluation in Spanish for people in Mexico. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 8/hour
  23. SPC

    Chatbot testing and evaluation

    United States · SpecialistChatbot testing featured on Alignerr's homepage. The public board also lists lower-paid audio work. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 40–120/hour
  24. RLH

    English writing

    United States · RLHFEnglish writing tasks. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 30–40/hour
  25. SPC

    Korean language

    Asia · SpecialistKorean language tasks. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 30–35/hour
  26. RLH

    Audio transcription

    United States · RLHFAudio transcription and transcript editing. This was most of the Alignerr board Parailabs recorded. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 10–35/hour
  27. SPC

    Healthcare QA, Central American Spanish

    Latin America · SpecialistHealthcare QA in Central American Spanish. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 15–20/hour
  28. COD

    Software evaluation

    United States · CodeSoftware evaluation. Handshake publishes a ceiling. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 150/hour
  29. RLH

    General AI evaluation

    United States · RLHFGeneral AI evaluation. Handshake publishes a ceiling. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.up to USD 17/hour
  30. RLH

    Entry tasks

    Named on the task · RLHFEntry tasks. Pay is per approved task; the hourly figure is an equivalent. The task names the country. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 15–30/hour
  31. RLH

    Generalist

    United States · RLHFOne Meridial generalist posting in the United States, at $20 an hour. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 20/hour
  32. COD

    Coding specialist

    United States · CodeMeridial coding specialist work in the United States. The rate depends on experience. Ranges are rechecked on 16 January 2027. The platform shows the rate before the work starts.USD 8–65/hour

32 of 32 roles · hourly, in USD

Filter roles

The talent

The team.

The people the desk matches to platforms: engineers, writers, and specialists, placed by region, language, and track.

  • Bogotá, Colombia (LATAM)

    Mateo

    Bogotá, Colombia (LATAM)Code evaluation
  • Manila, Philippines (Asia)

    Hanae

    Manila, Philippines (Asia)Multilingual rating
  • Kraków, Poland (Europe)

    Elise

    Kraków, Poland (Europe)RLHF and writing
  • Chicago, United States

    Amara

    Chicago, United StatesData and benchmarks

FAQ

Questions candidates ask.

Candidates

Who is this for?

People who can do the work AI-training platforms post: software, data, AI, writing, STEM, multilingual tasks, and other specialist tracks. Strong written English matters. Many projects look for two or more years of professional experience. The work is remote.

Still asking?

Write to the desk. A person answers, not a bot.

michael.allen@bridgewide.org

Contact the desk

All questions