AI Operator
Ground TruthCalibrate what you already know against what is actually true. Read any AI capability claim and know what it can and cannot do.
The AI workforce schoolEvery role is
Every role is
becoming an
AI role.
We train the people who run, shape, and build AI systems at work — operators, specialists, and architects, not just developers — with real depth where enterprises actually need it.
Start assessment →FREE · 5 MINUTES · INSTANT READOUT
AN ADVISOR FOLLOWS UP WITH YOUR ENTRY POINT
AN ADVISOR FOLLOWS UP WITH YOUR ENTRY POINT
Where our graduates leadLawrence Berkeley National LabP&GStandard CharteredFleetguardTata GroupInfosys
1000sOf mid-career professionals across dozens of live cohorts
4.8★Average graduate rating across cohorts
42Median age — built for mid-career, not bootcamp grads
01 · Why this school exists
In the 1980s, enterprise IT didn't stay inside computer-science departments — it created a workforce of operators and solution experts who ran the systems everyone else depended on. AI is doing it again.
Most AI courses train developers. But the people who will actually carry AI into organizations hold every other job title — and they need working depth, not survey slides.
02 · The diagnosisYou don't have an AI knowledge problem.
You don't have an AI knowledge problem.
You have a context problem.
Most senior professionals trying to learn AI fail in the same four ways. The pattern is so consistent we built the program around breaking it.
After eight years leading product, I'd rather sit through a board review than open another "Intro to LLMs" tutorial.
01≈ 14 hrs lost
The endless tutorial loop
Hours of YouTube on prompting. None of it survives Monday morning.
02≈ $800 wasted
The MOOC graveyard
Three courses started. Zero completed. Certificates aging on LinkedIn.
030 shipped
Tool-soup paralysis
Tabs open: ChatGPT, Claude, Copilot, Cursor, n8n. None integrated into your actual work.
04−5 hrs/wk
Stuck without a peer
No one to ask when the model is wrong, the API breaks, or the prompt won't generalize.
03 · The 100-day arcOne ladder. Six rungs.
One ladder. Six rungs.
Two doors at the top.
Six levels, about two hours a day for mid-career schedules. Everyone climbs the same spine — senior PMs, analysts, engineers, team leads. The only fork comes at Level VI.
AI Practitioner
The ToolkitDaily reps in Claude, ChatGPT, and Cursor. A reusable prompt and workflow library applied to your real work.
AI Builder
The FieldTake the toolkit into your day job, with weekly critique from a senior engineer. One AI initiative shipped inside your company.
AI Specialist
The CraftMove from delivery to depth — agents, RAG, evals, multi-agent. Pattern depth across the surfaces that matter in production.
AI Founder
The ForgeAuthor a complete AI system end-to-end with your pair — attested, replay-able, evaluated. A portfolio artifact you can defend in a board review.
AI Architect
The Charter · or · The ArchitectureStrategy Architect
Business depth — strategy, governance, transformation.
Solutions Architect
Production depth — MLOps, scaling, observability.
ONE SPINE FOR EVERYONE · CREDENTIAL AT EVERY RUNGMEDIAN AGE 42 · TEN-PLUS YEARS OF EXPERIENCE
Where you enter is personal — the assessment places you.Start assessment →FREE · 5 MINUTES · INSTANT READOUT
04 · How class worksLearner. Pair.
Learner. Pair.
AI. One build.
Every session runs as a trinity: you, a human pair, and an AI working the same problem. A learner alone with AI is a chatroom. A learner alone with a human is a tutorial. The triangle is the only structure adult learning at scale survives in.
Every hour is the same arc — three phases, twenty minutes each.
Phase 1 · 20 min
I Build
The AI coach builds in front of you, narrating the move. You watch, annotate, and ask — no keyboard yet. You see the move performed before you attempt it.
SurfaceTheater
Phase 2 · 20 min
We Build
You and your pair build it together. The AI is the third participant — referee, refresher, refactor. The Trinity, in full. Dialogue with a peer is where principles get extracted.
SurfacePair Canvas
Phase 3 · 20 min
You Build
You build it solo. The AI is on tap but silent unless asked. Your pair reviews on ship — the hand-off is the proof. Mastery accrues here, only here.
SurfaceSimulator mission
The same I-do / We-do / You-do model that built apprentice systems for a thousand years — rebuilt around the Trinity.
05 · The product, on the page
One minute of a real session.
No keynote. No marketing reel. Instructors and students working through a real problem on a real system — same room, same recording, unedited.
Eight minutes in. No edits, no overlays. This is the actual class, not a trailer.Surface · Theater · Live cohort excerpt
06 · Depth, not survey slides
The four things enterprises actually ask about.
Every track goes deep on the questions that decide whether AI ships at your company — the parts most courses skip.
01
Access control
Who can invoke which agent, with whose data, under which identity. Permissions as a first-class design problem.
02
Cost
Model tiers, token budgets, caching, and fallbacks — knowing what a workflow costs before finance asks.
03
Governance
Evals, guardrails, audit trails, and human-in-the-loop tiers that make AI systems defensible, not just demoable.
04
Data sovereignty
What leaves the building, what never can, and how to design AI systems that respect the boundary.
07 · The credentialOne fakeable signal.
One fakeable signal.
Two unfakeable ones.
MCQs can be gamed. Demos can be staged. Three independent signals, validated in sequence, cannot. We borrow the architecture medicine, aviation, and security all arrived at independently.
I · Transfer
Simulator mission
A novel enterprise scenario you've never seen. The model is reset. One shot to ship.
FAKE-RESISTANCE · VERY HIGHII · Breadth
MCQ bank
2,000+ items, IRT-calibrated. Tests coverage. The weakest of the three — and that's why it's not alone.
FAKE-RESISTANCE · LOWIII · Sustained
Kapi project · pair
A real system, deployed, demoed to your pair, signed off. Six weeks that has to actually run.
FAKE-RESISTANCE · VERY HIGHThe same architectural answer medicine, aviation, and security all converged on independently. Not novel pedagogy. Applied pedagogy.
08 · PitstopNever stuck for more
Never stuck for more
than an hour.
MOOC failure is rarely a curriculum problem — it's a momentum problem. A real engineer answers on Pitstop within the hour, including weekends. The Pair holds you to the rhythm.
"I was stuck on a retrieval pipeline at 11 pm on a Sunday. An engineer was on a call with me by 11:42. That's when I knew this was different."
< 1hPitstop response target
24/7Including weekends
3×Trinity: learner · pair · AI
75+Live batches taught
09 · FacultyWe're not just teachers.
We're not just teachers.
We're practitioners.
The team behind Mitra — India's first social robot. Every instructor has shipped real AI systems into production. We teach what we've actually deployed, and we still deploy.
BV
Dr. Balaji Viswanathan
Co-founder & CEOEx-Microsoft. Led the creation of Mitra, India's first social robot, deployed at major airports and enterprise events. Most-followed person on Quora with over a million followers.
MR
Mahalakshmi Radhakrishnan
Co-founder & CTORoboticist and engineering leader. Co-created the Mitra robot's intelligence layer. Has shipped human-robot interaction systems into production enterprise environments.
Plus guest practitioners from
AnthropicMicrosoftIBMStripe
10 · ApplyFind your level
Find your level
in five minutes.
A short, honest assessment that places you on the right rung of the ladder — L1 to L6. Rolling enrollment, pair-matched on signup. No GMAT, no essays.
Start assessment →FREE · 5 MINUTES · INSTANT READOUT
Prefer a human first? Use the form — an admissions lead replies within two business days.