Practice before theory
Theory is introduced exactly when you need it — that's when it sticks.
a profession simulator
LLM Operator
No code required. Turn any wish into a task the model can actually do — and take real work off your plate this week.
$299· from $12/moFree · Ch 18 chapters · ~3.8h
~35 min · works with any model runnerChapter 1 is free. You only pay when you decide to continue.
Task 2 of 4
Sasha pokes her head in: "So, figured out what's wrong?" Before answering her, break Slava's request down into its parts.
Slava's request was: "Make a nice post about our new pricing plan." Which elements of a real task are missing from it?
AOnly the audience — the rest is fine
BOnly the channel — Telegram or Instagram, it's not specified
CAll four: audience, channel, constraint, and action
DNothing important — the model should have just asked Slava for details
Auto-check: correct
WhyThe four elements aren't independent — they constrain each other.
This isn't an online course with lectures — it's a simulator of real work.
80% practice · 20% theory×2 your hourly rate
- Chapter 1 free forever
- No card required
- Cancel any time
- Progress is saved
2 people are already taking this
Approach
Why it works
Verifiable certificate
Proves your skills with a publicly checkable link — add it straight to your résumé.
Real market data
Salaries: remote-market medians and hh.ru / Habr Career — both markets shown.
Progress & streaks
Persist between sessions to keep you on track.
Flexible payment
Bank card · cryptocurrency.
The super-profession
LLM Operator is the base literacy of the AI era — the one skill every role needs regardless of title. You stop treating the model like a slot machine and start operating it like a tool with inputs, checks, and a done-criterion.
What you'll be able to close on your own
- Turn any vague request into a task with a checkable done-criterion before writing a single prompt
- Pack the model with exactly the context it needs — style samples, facts — without leaking what shouldn't be there
- Fix a near-right answer with one targeted edit instead of regenerating from scratch
- Catch a confident, plausible-sounding fabrication by checking it against the source
- Order the same result in three different formats for three different readers
- Run one task across a batch of objects with a template and a spot-check
- Assemble a personal weekly pipeline that turns a real recurring task into a repeatable process
The AI-mastery scale
Where are you on the AI-mastery scale?
Most people never get past asking ChatGPT a question now and then. The top levels — automating your work and building AI systems — are rare, and that's exactly what makes them valuable. This is the band the school lifts you into.
Click a level to see the skills it unlocks
Percentages are an illustrative estimate of the market distribution, not hard data.
The problem this solves
Most people given AI access use it like a slot machine: type a vague wish, get something vague back, regenerate, hope. Nobody taught them the model can only be as precise as the request — so they lose hours to 'not quite right, try again' instead of minutes to a task with a clear done-criterion.
What changes for you
You stop gambling on regenerations and start operating the model like a tool: a task with a checkable done-criterion in, a verified result out. It's the difference between hoping and knowing — and it shows up in your very first artifact, built on your own real work.
Who it's for
Anyone just handed AI access at work with no training
Stop losing an hour to regeneration roulette. Get a task-spec formula you use before every request, starting today.
Marketing, sales, ops and HR specialists writing content daily
Order the right format for the right reader on the first try — table for your boss, message for the team, letter for the client.
Managers who want their team using AI reliably, not randomly
Get a shared vocabulary — task-spec, context-pack, verify — you can hand to anyone on the team.
Anyone considering a deeper AI-profession track
Find out, in one free chapter, which of the three specialist tracks fits how you actually think — before paying for one.
Who it's not for
- People who already have a disciplined prompting workflow — this teaches the fundamentals, not advanced technique.
- Anyone looking for prompt templates to copy-paste — you'll build judgment, not snippets.
- Developers who want to build AI systems — that's Automation AI Engineer, not this.
How the learning works
It's a simulator, not a course
No lectures about AI. You inherit a real workplace scene on minute one — a coworker's failed chat session — and make the call yourself.
Your own work is the material
Chapter 1's key task isn't a made-up exercise — it's a real task from your own week, formatted through the formula you just learned.
Answer first, theory after
You commit to a diagnosis before any explanation appears. Theory is revealed only after you've made the call — that's what makes it stick.
No code, ever
Every task happens inside a normal chat with a model. If you can type a message, you can do this course.
What you'll learn
Ask like an operator
- Turn a wish into a task with a checkable done-criterion
- Pack the model with exactly the context it needs
- Fix an answer with one targeted edit, not a full regeneration
Trust, but verify
- Catch a confident fabrication by checking the source
- Order the right output format for the actual reader
- Run one task across a batch with a spot-check policy
The career this opens
Where the profession is in demand
Examples of companies hiring AI engineers. Not affiliated with the school; employment is not guaranteed.
8 chapters · ~3.8h
- Free
A wish is not a task
35 minWatch Slava lose an hour to regeneration roulette. Learn the one move that ends it.
- Locked
Context decides
25 minA client letter that doesn't sound like the client. Learn what to hand the model — and what to withhold.
- Locked
Fix it, don't reroll it
25 minSlava is on his fourth regeneration and losing the good parts each time. Learn the targeted edit instead.
- Locked
Verify, then trust
25 minA confident, well-written answer with a fabricated number inside. Learn to check the source, not the vibe.
- Locked
Format for the reader
25 minOne wall of text sent to three different people. Learn to order the right shape for the actual reader.
- Locked
Run it as a batch
25 min18 personalized letters, one at a time, by hand. Learn the template-plus-spot-check move.
- Locked
Your own pipeline
30 minPick one real weekly task and turn it into a repeatable five-step process.
- Locked
Capstone — Monday's report
40 min40 minutes to assemble the real deliverable: your pipeline's output, plus how you catch its mistakes.
Who teaches this
Alexander Mihalkevich
Founder, mihalkevich.com
I build and ship products solo with AI agents every day. This course is the exact discipline I run on — no theory I don't use myself.
mihalkevich.com →Why now
AI access rolled out to entire teams in the last year, but almost nobody was taught how to operate it — most people are still regenerating and hoping. The gap between 'has access' and 'gets reliable results' is where this course sits, and it closes fast once operating discipline becomes common knowledge.
Pick how you learn
Start with Chapter 1 free. Pay only when you decide to continue.
Самостоятельно
from $12/mo
$299 one-time · or 24 payments
- Все главы курса
- Авто-проверка заданий
- Сертификат о прохождении
С ментором
Best valuefrom $21/mo
$499 one-time · or 24 payments
- Всё из «Самостоятельно»
- Проверка заданий ментором
- Личная обратная связь
- Чат когорты
С трудоустройством
from $33/mo
$799 one-time · or 24 payments
- Всё из «С ментором»
- Резюме и портфолио под конкретные вакансии
- Тренировка собеседований с разбором
- Не нашёл работу за 6 месяцев — вернём деньги
Payback: The first artifact — a task-spec template applied to your real work — ships inside the first 15 minutes of Chapter 1. By the end of the full track you walk away with a personal AI pipeline for one real weekly task, timed before and after.
- 14-day refund if it's not a fit
- Employment tier — on written terms
- 1-on-1 mentor review
- Lifetime access
Launch pricing · seats in this cohort are limitedPrice rises after the cohort fills
Join the community on Telegram
New chapters and simulators, breakdowns of course tasks and discussion with other learners. Free.
What stays in your portfolio
- A personal task-spec template you can reuse on any request to a model
- A verification checklist for your own routine
- A format playbook for the people you write for
- A batch template with a spot-check policy
- A working weekly pipeline for one real recurring task, timed before/after
The certificate you'll earn
A real, shareable credential — issued the moment you complete the chapter. Add it to LinkedIn or send the link.
The course certificate is issued once your final GitHub project passes the automatic checks — the project is mandatory.
Still on the fence?
Chapter 1 is free and self-paced — no card, no commitment. If the first hour doesn't feel like the real job, you've lost nothing. If it does, you'll know exactly why the rest is worth it. The only way to find out is to do the first chapter.
Start Chapter 1 free →FAQ
LLM Operator
Your first day at Yavno agency. Everyone just got AI access and was told to 'work faster'. Slava has been regenerating the same post for an hour. Sasha, the ops director, hands you his chat window: 'Watch. In 15 minutes tell me what's wrong.'
Start free