About rustLabs

Real experts, in the loop.

rustLabs is an applied AI lab. We build the environments, rubrics, and evaluations that AI labs train and test their models against — and we run rustLabs Experts, the network of more than 5,000 working professionals who supply the human judgment behind that data.

experts with rustLabs: 5,000+
expert fields: 6
per hour, by expertise: $20–100+
remote, on your schedule: 100%

01 · Why we exist

Models learn from judgment, not just text.

Modern models don’t get better on scraped text alone. They get better when an accountant walks through a return line by line, when an engineer explains exactly why a fix is wrong, when a clinician grades a diagnosis step by step. Our job is to capture that judgment carefully enough that a model can actually learn from it.

We’re a small team, and we like it that way. We work directly with the labs training models and directly with the experts doing the tasks — no layers of vendors in between. When a project needs a new environment, we build it. When an expert flags a broken task, a person reads it and fixes it, usually the same day. That’s the whole pitch: real experts, real oversight, and data we’re willing to put our name on.

02 · What we build

Three things, done properly.

01RL environments & evalsTask environments, grading rubrics, and evaluation suites designed around how labs actually train and test their models — not generic annotation templates.
02The experts platformOur own platform, built in-house. Experts apply, qualify, pick up projects, submit work, and get paid by the hour — all in one place, with a human on the other end.
03Partner data programsOngoing programs for AI labs and applied-AI companies, from small pilots to production pipelines — like our computer-vision work with ViaTouch's VICKI retail machines.

03 · How we work

The standards we hold ourselves to.

01Experts come firstClear task specs, fair rates, and fast answers. The data is only as good as the people making it, so we treat them like the professionals they are.
02Quality over volumeEvery submission is reviewed. We'd rather ship a smaller set a lab can trust than a mountain of noise — and our partners keep coming back because of it.
03Small and directYou talk to the people actually building the environments and reviewing the work. No vendor layers, no ticket queues, no runaround.

04 · Talk to us

Building something that needs experts?

Tell us what you’re working on and a researcher on our team — a person, not a queue — will get back to you.

Experts: apply to rustLabs here.

05 · Follow the work

See what we’re shipping.

New fields opening for experts, partner launches, and notes from the work itself — we post it as it happens.

Models learn from judgment, not just text.

Three things, done properly.

The standards we hold ourselves to.

Building something that needs experts?

Talk to a researcher

See what we’re shipping.