Aced Talent Network

Now accepting applications

Get paid to teach AI what you know.

Frontier AI labs pay Aced-vetted engineers and domain experts to build the benchmarks, evals, and training data behind the next generation of models. Remote, flexible, and paid at expert rates — part-time or full-time.

Apply to join How it works

$50–150+

Typical hourly rates

5–40 hrs

Per week, you choose

100%

Remote

Paid

Skills screen

The work

Knowledge work, not labeling.

Labs don't need clicks — they need judgment. Four kinds of projects flow through the network, all remote and scoped with rates posted up front.

Benchmarks & evals

$60–150/hr

Author the tasks, test cases, and gold answers that measure what frontier models can and can't do in your field.

Python SWEMLDomain experts

Model-in-the-loop data

$80–200/hr

Work alongside lab engineers in real time — reviewing model outputs, breaking edge cases, and shaping training data.

Senior SWEMedicineLawFinance

Training data & preferences

$50–120/hr

Write, rank, and refine the expert demonstrations models learn from — the highest-leverage writing you'll ever do.

All domainsWritingScience

Rubric design & QA

$70–150/hr

Define what "good" means: build grading rubrics, calibrate reviewers, and audit data quality for entire projects.

Senior ICsManagersEducators

How it works

Apply once. Get matched as demand comes in.

Apply in 5 minutes

Complete the job matching profile on your Aced account — domain, availability, and rate. No cover letter.

Take a paid skills screen

A ~45-minute work sample from your track, graded against the same rubrics you'd use on the job. You're paid whether or not you pass.

Get matched to projects

When a partner lab needs your profile, we reach out with scope, rate, and timeline. Accept what fits, skip what doesn't.

Quality

A talent pool that stays sharp.

Proven skills, not resumes

Every member passes a paid, rubric-graded work sample in their track before they see a single project.

Track record you can't fake

Aced members bring years of interview practice, peer mocks, and community history — signal most talent pools never see.

Quality that compounds

Every task is scored. Top contributors get first access to the best projects; quality below the bar means leaving the pool.

Apply

Join the network.

Five minutes now, a paid skills screen later. Your name stays private — partners see an anonymized profile until you accept an introduction.

Applying takes a free Aced account — your practice history and peer reviews strengthen your application from day one.

FAQ

Questions, answered.

AI labs need experts to create the material models learn from: benchmark tasks, evaluation rubrics, expert demonstrations, and reviews of model outputs. It's knowledge work — closer to teaching and editing than labeling.

Frontier AI labs and enterprise AI teams. Specific names are shared after you pass the skills screen and sign an NDA — confidentiality is part of the job.

Rates depend on domain and seniority. Most software engineers land between $60 and $150 per hour; specialists like physicians and attorneys can earn more. Every project posts its rate up front, and the skills screen itself is paid.

Most members contribute 5–20 hours a week alongside a full-time job. Some projects offer 35+ hour engagements for members who want full-time work.

No. Labs want deep expertise in your field — we train you on the tools, data formats, and rubric-writing before your first project.

Usually, yes — engagements are remote and scheduled around you. Check your employer's moonlighting policy; work for a direct competitor of your employer is off the table.

Three layers: a structured application, a paid rubric-graded skills screen in your track, and ongoing per-task quality scores once you're matched. This is how we keep the pool strong enough that labs come back.