Skip to content
AI Training Pay
esc
  • Loading roles…
  • Pages and tools

OpenTrainPhDMaster's

ML Engineer, Machine Learning Training and Inference

You apply here in about two minutes and finish on the hiring platform. The next step is a short AI interview (technical roles include a coding challenge). Passing it is how you qualify for paid tasks, so take it soon after you are invited.

Field
Machine learning
Where
United States, Algeria, Argentina, Austria, Bahrain, Bangladesh +46 more
Commitment
part time
Hours
up to 15/week
Asks for
PhD, Master's
Posted
5 days ago

Pay data for your profession: Data scientist AI trainer jobs · PhD AI trainer jobs

What the listing asks for

Hours
About 15 a week“Schedule: Flexible contractor work of approximately 15 hours per week, with less than 20 hours per week expected.”

Read from OpenTrain AI's listing text; only what it states. Full listing

How this pay compares

At $90/hr (the midpoint of $60 to $120), this role pays more than 45% of the 68 open machine learning roles with a listed pay range, the same as the median.

Open machine learning roles that ask for a Master's have a median of $106/hr (12 roles).

Among platforms with 10 or more priced machine learning roles, OpenTrain AI has the highest median ($109/hr across 34 roles) and Alignerr the lowest ($63/hr across 13).

OpenTrain AI roles that closed since late October stayed listed a median of 26 days (289 roles). OpenTrain AI posted this one 5 days ago.

Computed from 3,087 open roles we track across Mercor, micro1, Handshake AI, Alignerr, Outlier, Turing, Appen, OpenTrain AI and OneForma, refreshed daily. Midpoints of listed ranges; “up to” ceilings are left out of medians.

What you would earn at this rate

Per week, before tax

$1,350

Per month, if the work is steady

$5,850

Tax
Contract pay arrives with no tax withheld. A US contractor at this rate and these hours would set aside about $1,479 a month for self-employment tax plus federal income tax at the 12% bracket (25%, before state tax). Outside the US, your own country's rules apply.

Assumes paid hours only; unpaid assessments, onboarding and gaps between projects lower the real figure. Full take-home calculator · Track what you actually earn

Open machine learning roles on each platform

The same kind of work, priced on every platform we track, with the best-paid open role on each.

“Up to” rows list only ceilings, so their median is a maximum, not comparable with the range midpoints above them.

All 128 open machine learning roles →

The full listing

As posted on OpenTrain AI. See the original

The work

  • Create, solve, review, and validate machine-learning engineering tasks.
  • Develop and validate machine-learning models, training pipelines, inference systems, and supporting infrastructure.
  • Implement model components, data pipelines, evaluation systems, and numerical methods.
  • Build reproducible workflows with Python and command-line tools.
  • Optimize memory use and throughput, and debug numerical or system-level failures.
  • Assess implementations against objective correctness and performance requirements.

What it pays and takes

  • Pay: $60 to $120 per hour equivalent, paid per task that meets the project specifications.
  • Schedule: Flexible contractor work of approximately 15 hours per week, with less than 20 hours per week expected.
  • Location: Fully remote and available in the countries listed for this role, subject to applicable hiring eligibility.
  • Language: Professional fluency in English.
  • Education: A BSc, MSc, or PhD in computer science, machine learning, artificial intelligence, applied mathematics, statistics, engineering, or a closely related quantitative discipline.
  • Skills: Practical proficiency with Python and coding agents, plus the ability to reason about model behavior, training and inference workflows, numerical methods, reproducibility, and performance.
  • Useful tools include PyTorch, JAX, NumPy/SciPy, SGLang, vLLM, llama.cpp, tokenizers, and transformers.
  • Experience in an established technology company, AI laboratory, research organization, open-source project, or academic setting is preferred.

How it works

You apply here in about two minutes and finish on the hiring platform. The next step is a short AI interview (technical roles include a coding challenge). Passing it is how you qualify for paid tasks, so take it soon after you are invited.

The hiring platform's application for this job asks for your resume, your LinkedIn profile, your phone number, your location and answers to 8 screening questions.

About AI training work

AI training work is the human work behind modern AI systems, including building datasets, testing models, and reviewing how systems perform. People with strong technical skills are paid to help make these systems more accurate, reliable, and useful.

From applying to your first payment on OpenTrain AI

  1. 1

    Create a profile

    Create a free OpenTrain profile with your domains, languages and hourly rate.

  2. 2

    Apply

    Apply to the listing. Some listings hire through OpenTrain; others send you to the hiring platform.

  3. 3

    Funded milestones

    The client funds each milestone before work starts.

  4. 4

    First payment

    OpenTrain pays out to your connected payout method when the client approves the milestone. It doesn't publish a fixed payout schedule.

Sources: OpenTrain: AI training careers, OpenTrain docs: invoicing and payouts

Full OpenTrain AI profile: watch-outs and the public record →

Pay history

OpenTrain AI has listed this role at $60-120/hr since Nov 18, 2026 with no change.

Across OpenTrain AI's machine learning roles, the median listed pay was $90/hr on Nov 1, 2026 (23 roles) and is $109/hr as of Nov 22, 2026 (34 roles).

What workers report about OpenTrain AI

Anonymous reports from people who applied or worked there: whether they got work, what they were paid and how fast.

Other machine learning roles

All 128 machine learning roles →

Get new machine learning roles by email

New roles fill fast. We email when a match is posted, at most once a day. Unsubscribe in one click.