You spent a career becoming an expert. The platforms still reject 4 out of 5 experts who apply — because nobody taught them rubric craft, golden answers, or what an auditor actually flags. We do.
This school wasn't built by marketers. It was built by working SMEs who learned the craft the hard way, inside real frontier-lab projects — and have the receipts.
The programs teach exactly what we practice: the frameworks we use on live projects, the errors we see as auditors, and the standards our own work is held to.
One of our founding instructors earned more than $400K in twelve months of expert AI-data work — task creation, review, and auditing. Documented and verifiable; evidence available under NDA to enrolled students and partners.
This is one practitioner's documented result, not a promise of income. Your results will vary with credentials, effort, and market demand. See full disclaimer.
Design expert-level tasks that legitimately challenge frontier models — realistic context, cascading dependencies, objective gradability.
Write self-contained, weighted criteria — including the negative criteria and criticality discipline most experts have never heard of.
Produce reference answers with verified citations and complete source engagement — the anchor of every task package.
Evaluate model responses in the strict formats projects demand: excerpt-anchored, present-tense, one error per item, calibrated.
Learn what reviewers actually flag, survive QC with your ego intact, and climb to the reviewer/auditor tiers where pay is higher and queues are fuller.
Platform strategy, portfolio assembly, rate negotiation with real market data, and sustainable pacing — the anti-burnout module.
If a platform has turned you down despite your credentials, this is almost certainly why.
Every program here is priced in hours of the work it opens up. Published expert-tier rates for credentialed professionals run roughly $75–$200 an hour, with ultra-specialist rates reported well above that. Here is what each program costs at those rates.
About three hours at $150/hr — and about five at a conservative $100/hr. An afternoon of the work it prepares you for.
The conservative case. Even at the bottom of the published band, it is under a week of part-time work.
If you bill $400 an hour, full access costs a little over one billable hour — and unlike a CLE credit, it does not expire.
One of our founding instructors earned more than $400,000 across a single year of expert AI-data work — task creation, review and auditing. Documented and verifiable; evidence available under NDA to enrolled students and partners.
Now the part most people selling you something leave out. That is one senior practitioner working at high volume in a strong market. It is not typical, and we will not pretend it is. This is contract work: volume fluctuates, rates vary by domain and platform, and a quiet quarter is a real thing. We have no student outcome data yet, because we have none worth publishing — when we do, it goes up, including the results that do not flatter us. What we will say plainly is this: the work exists, it pays well, it is growing fast, and the reason most credentialed professionals never reach it is a craft gap that takes four weeks to close.
Earnings disclosure. The figure above is the documented result of one specific individual under specific market conditions, provided as factual background. It is not a promise, projection, guarantee, or representation of typical or expected results. No income, placement, platform acceptance, or engagement volume is guaranteed. Your outcome depends on your credentials, your available hours, the quality of your work, and market demand — all of which vary and are outside our control. Rate ranges cited are drawn from public platform reporting and change over time. See our full disclaimer.
There is no ladder to climb and no tier held back. Everything we teach sits behind a single paywall, and $495 opens all of it — permanently.
14-day refund, no argument. See Terms.
Splitting prompts from rubrics from audit would sell you a third of a skill. The work is judged whole, so it is taught whole.
Nothing here depends on a date, a cohort filling, or you being free on a Tuesday evening. You start when you start.
At published expert-tier rates, $495 is roughly three hours of the work it prepares you for — and one hour of your own billable time if you charge like a partner.
Graded cohorts and a reviewer intensive are on the roadmap, and people who own the Handbook will hear about them first. We are not taking money for them today, and we are not going to pre-sell a date we have not earned.
No. The programs are designed for credentialed professionals with zero technical background. If you can write a memo or a workpaper, you can learn this craft.
All of them — the craft is transferable. We are vendor-neutral and teach the skills every serious platform screens for. We never share any platform's assessment content or coach you to game a specific test; that's in our published Integrity Code, and it's why our training holds up.
Yes — it is the documented earnings of one of our founding instructors across twelve months of expert AI-data work, verifiable under NDA. It is also not a promise: it reflects senior credentials, high output, and strong market demand in that period. We publish honest market data, including the downsides.
The same skills — rubric design, evaluation, QC — are exactly what enterprises are hiring for as they build internal AI evaluation and governance functions. We teach AI evaluation competence, not platform tricks.
Immediately. Access is not tied to a date, a cohort filling, or a timetable — you pay, everything unlocks, and you work through it at whatever pace your practice allows. Live graded programs are on the roadmap and Handbook owners will hear first, but we are not pre-selling a date we have not earned.
We reply within one business day, and a credentialed practitioner answers — not a support queue. Tell us your background and what you are trying to get into.