Process
Five stages, in this order.
Whether the contract is yours or ours. The pilot is paid, priced at the volume rate, and the output is yours either way.
The stages
From first call to a batch shipping weekly.
- 01
Scope
One call with whoever owns the data. We read your spec, your rubric and a sample batch, then tell you which parts we can staff this month and which we cannot. The second answer is the useful one.
- 02
Pilot
A small paid batch at the full quality bar, in whichever system you chose, through your QA gate. You keep the output whether or not it continues, and the price is the price it would be at volume.
- 03
Staff
Named experts, not a queue. You see the bench before it starts: who they are, what they have practised, how long, and who leads the squad. You can reject anyone without giving a reason.
- 04
Run
Fixed weekly throughput, one point of contact, and daily visibility into what shipped. Sampling and double-scoring run inside every batch rather than after it, so a drift shows up in days rather than at the end of the month.
- 05
Audit
Every batch ships with its agreement statistics, its reject rate, and the items we threw away and why. If a number moves in the wrong direction, you hear it from us before you find it.
Quality control
Two things, and they are both boring.
Sampling: a fixed percentage of every batch is re-scored blind by a second expert who did not see the first answer. Calibration: the squad re-runs a shared set at intervals, so drift shows up as a number rather than as a complaint from your side.
Both figures ship with the batch. So does the reject rate and the list of items we threw away. A vendor who only reports throughput is reporting the easy half.
What we need from you
A spec, a sample, and someone who answers.
A spec, a rubric if you have one, and a sample of what good looks like. If you do not have a rubric we will write one and you will own it, but expect the first draft to be wrong in ways only your researchers can see.
Beyond that, one named person who can answer questions about edge cases within a day. Expert data stalls on unanswered judgement calls more than on anything else.
Pricing
The pilot price is the volume price.
Per item for supervised demonstrations and reinforcement learning and rubrics, per hour for computer-use trajectories and evaluation, and a fixed fee for an environment build.
Rates depend on the domain and the licensing required. A practising physician on a US-licensed project and a bookkeeper are not the same line item, and we will not average them into one.
Tell us the domain you are stuck staffing.
One call, and you get a straight answer about whether we can take it. If we cannot, we will say so on the call rather than three batches in.
Prefer to read first? See who runs Xocket Data Labs or every way to reach us.