Accepting applications

Get paid to train
frontier AI

Join thousands of experts earning $40–$150/hr reviewing AI outputs and labeling data — on your own schedule.

$40–$150/hrexpert pay/190,000+annotators/$36M+paid out/50+task domains
chrisqing — task.sh

$ chrisqing task fetch --field=coding

{

task_id: "an_8f21c3",

category: "coding",

rate: "$75–$150/hr",

reviewers_active: 186,

status: "available",

}

$

infra.status

99.9% pipeline uptime<24h review turnaround50+ eval domainsPhD-reviewed quality gate

190,000+

Active annotators

$36M+

Paid out

50+

Task categories

5

Countries

Pipeline

Four stages, one pipeline

Every trainer runs the same pipeline — init, eval, verify, ship — usually inside 24 hours.

[01]

$ init --profile

Create your profile

Create your account with email or Google. Add your profile, country, LinkedIn, and choose your expert categories.

[02]

$ eval --category=all

Pass the eval

Take a short assessment in your chosen categories. Demonstrate your expertise and unlock the next step.

[03]

$ kyc --verify

Verify identity

Complete identity verification (KYC). Once approved, you get full access to the dashboard and paid tasks.

[04]

$ payout --start

Ship tasks, get paid

Complete annotation tasks at $40–$150/hr and withdraw your earnings directly to your crypto wallet.

Tasks

Three ways to train the model

Every pipeline reduces to one of three calls. Your expertise decides what goes in.

rank_responses(prompt, candidates)

// signal

Comparing model responses teaches the AI which answers are actually correct, helpful, and safe.

// spec

Read a prompt and two or more model answers, then rank them and explain what makes the best one better.

review.py

# Rank these responses

  • [x]Is the answer factually correct?
  • [ ]Does it fully address the question?
  • [ ]Is the reasoning sound, not just the result?
  • [ ]Would an expert in the field approve it?
Why this pipeline

Built to train models, not waste time

Every part of the pipeline assumes you're a specialist whose time is expensive — because it is.

Priced like engineering work

$40–$150/hr. Reviewing model outputs in STEM, coding, and specialized domains is compensated like the skilled work it is.

Async by default

No shifts, no standups, no manager. Pull tasks from the queue whenever it suits you, push whenever you're done.

Fast pipeline, clear status

init → eval → verify → ship. Most trainers clear the full pipeline and reach paid tasks inside 24 hours.

On-chain payouts

Withdraw straight to your own wallet on Ethereum, Polygon, and more. Balance state is always visible, never a black box.

You're training the model, not busywork

Every submission is a training signal — proofs, code reviews, clinical judgment — that directly shapes model reasoning.

Reviewed pipeline, verified reviewers

Every trainer clears an eval and an identity check before shipping. 190,000+ verified specialists run this pipeline.

Active pipelines

Live training runs

A live snapshot of pipelines currently accepting reviewers.

View all runs
#mathematics$60–$120/hr

Advanced math proof review

AKMPJL
240 reviewing now
Join run
#coding$75–$150/hr

Python code completion rating

TSRNDC
186 reviewing now
Join run
#medicine$75–$150/hr

Clinical text evaluation

ALHWKB
95 reviewing now
Join run
#finance$75–$150/hr

Financial reasoning tasks

VNGBPS
132 reviewing now
Join run
#law$75–$150/hr

Legal document analysis

MRECTF
64 reviewing now
Join run
#physics$60–$120/hr

Physics problem solving

OKLMRB
118 reviewing now
Join run
domains.json

Pick a training domain

Rate scales with difficulty — routine eval at the low end, frontier reasoning at the top. Most trainers specialize in 1–2 domains and stay for years.

View domains.json

$60–$120/hr

> Math

$75–$150/hr

> Programming

$75–$150/hr

> Medical

$75–$150/hr

> Law

$75–$150/hr

> Finance

$60–$120/hr

> Physics

Found your domain? The eval takes ~30 minutes — every submission is reviewed by hand.

Start the eval
eligibility.json

If you're a specialist, there's a queue for you

ChrisQing routes real expertise into model training. You bring the domain depth; the pipeline brings the tasks, the rate, and the schedule.

  • Based in the US, UK, CA, AU, or NZ
  • Degree-level or equivalent hands-on depth in your field
  • Able to clear a domain eval on the first or second attempt

Grad students & researchers

Route your coursework and research depth into a paid pipeline between classes and projects.

Domain professionals

Doctors, lawyers, and finance experts feed real-world judgment into high-value training runs.

Engineers & data scientists

Review model code, audit reasoning traces, and catch regressions at rates above most freelance work.

Subject-matter experts

Any field with real depth — linguistics, chemistry, whatever — has an open queue waiting for it.

payouts.log

Ship tasks, earn $40–$150/hr

Top reviewers clear thousands monthly. Depth in STEM, coding, and specialized domains is priced accordingly.

$ whoami
"Six months on the pipeline and it's never gone dry. Async, well-paid, and I can see exactly what my reviews change."

@sarah_k

math + physics · pipeline reviewer

$12,400

verified
$ whoami
"The coding eval tasks are basically code review with better pay. It's the best-paid freelance-adjacent work I've found."

@james_r

coding + cs · pipeline reviewer

$18,200

verified
$ whoami
"The eval was fast and unambiguous. Within a week I was reviewing clinical data at the top of the rate band."

@dr_amira_l

biology + medicine · pipeline reviewer

$9,800

verified

$ balance --live

$3,420.00

review::python+$180.00
eval::math_proof+$240.00
withdraw::usdc−$1,000.00
ledger.log

Get paid on-chain, per submission

No invoicing, no net-30. The ledger updates as you work; withdraw whenever you're ready.

  1. 01Earn per submission

    Every merged submission adds to your balance at $40–$150/hr, priced by category and difficulty.

  2. 02Live balance, no black box

    Your dashboard streams exactly what you've earned and what's withdrawable, in real time.

  3. 03Withdraw on-chain

    Cash out to your own wallet on Ethereum, USDC, or USDT once you clear the minimum — on your schedule.

docs.md

FAQ / docs

We currently accept trainers based in the United States, United Kingdom, Canada, Australia, and New Zealand. You need to clear a domain eval and pass identity verification.

$40–$150/hr depending on domain and task difficulty. Specialized queues — medicine, coding, advanced math — sit at the top of the band.

After signup you pick your domains and run a short multiple-choice eval. Pass it, clear identity verification, and the task queue unlocks.

Payouts are on-chain. Once your balance clears the minimum, withdraw to your own wallet on Ethereum, Bitcoin, or USDT.

Most trainers clear the full pipeline — signup, eval, KYC — inside 24 hours. Some clear it in under an hour.

Math, coding, biology, finance, creative writing, law, and more. Tasks are matched to your verified domains automatically.

Ready to ship your first task?

190,000+ trainers are already on the pipeline. Async hours, on-chain payouts, work that ships.