AI Fundamentals Mastery
Learn how AI actually works the AI-driven way — twelve small things that really run, from catching a chat model confidently inventing an answer to shipping your own grounded, tool-using, evaluated agent that wakes up on a schedule.
✓ First 2 lessons free · no card needed · one plan unlocks every course
The course at a glance
A from-the-ground-up course for anyone who uses AI and wants to understand it — and for working professionals who suspect there are gaps in what they think they know. One metaphor carries the whole way: the model is a brilliant new hire who read the whole internet and remembers nothing about you. Everything after that is onboarding — the job description, the desk, the filing cabinet, the tool belt, the approval chain, the performance review.
Every one of the twelve phases ends in a small thing you finish and can show — a Guess Machine, a Token Meter, a Prompt Lab, Chat With Your Docs, a Tool Belt, an Inbox Triager, an Eval Harness — until the final phase, where they compound into The Apprentice: an agent that answers from your own documents with citations, uses tools, remembers across sessions, stops at you before anything irreversible, and passes a fifty-case eval suite before it ships.
How learning happens
Every chapter follows the same loop — taught in Phase 0 and run for the rest of the course. Two beats to learn, two beats to build. Most chapters take 30–60 minutes; the build chapter that closes each phase runs longer.
Learn
A short read plus an animated model of the idea. No syntax dumps — you learn what the thing is for.
Brief
Write the instruction in plain English. Precision of intent is the skill being trained here.
Review
Read what came back against a checklist. This is where engineers are actually made.
Tweak
Change one thing yourself, by hand, and watch it move. Ownership beats copy-paste.
What you need first
A browser and a free AI account
Phases 1–5 need nothing more than a chat window and somewhere to keep notes. No code, no API keys, no jargon to start.
Python and an API key
The engineering half runs real Python on your machine — but you never write it from a blank page. You brief the AI, read what it wrote, and fix it.
Twelve finished mini-projects
Machine → tokens → prompts → working agreement → context → RAG → retrieval → tools → agents → patterns → evals → The Apprentice.
The portfolio
Everything built in this course is real, runnable, and shareable — and it compounds into one flagship artefact.
Understand and direct it
The Guess Machine, the Token Meter, the Prompt Lab and your Working Agreement — what an LLM is, what it costs, and how to direct it like an editor, not a fan.
Ground it, equip it, loop it
The Long Conversation, Chat With Your Docs, the Retrieval Bake-Off, the Tool Belt, the Inbox Triager and the Router — context, retrieval, tools, agents and patterns.
The Eval Harness + The Apprentice
Fifty golden cases and a calibrated judge — then every earlier piece assembled into one supervised agent that runs on a schedule.
The capstone
Phase 12 is the reason the previous eleven exist. Every earlier project becomes one layer of a single agent — and every line of its configuration traces back to the phase that taught it.
The Apprentice
Answer from truth, not memory
Every answer comes from your own documents via hybrid search and reranking, with a citation attached — or an honest refusal.
Give it hands
Three real tools with typed schemas and honest descriptions — the one that can change the world sits behind an approval gate.
Let it choose the path
An outer plan with inner reason-act-observe loops, memory that survives a restart, and compaction when the desk fills.
Measure it, then gate on it
Fifty golden cases, a calibrated judge, and a threshold that blocks the ship when the number is red.
Concept coverage matrix
Every major concept, where it's introduced, and where it pays off.
The skill ladder
What you can honestly say after each stage. Each claim maps to something you've built, run, and can show on a screen.
I understand the machine
I can explain an LLM without "magic", estimate a job's cost, and A/B test prompts with a scoreboard.
I direct and ground AI
A working agreement, a repaired long conversation, and a docs chatbot that cites its sources.
I build agents
Typed tools, an agent that loops to done and stops at me, and a router with measured cost.
I ship real AI systems
A calibrated eval suite, and The Apprentice — grounded, equipped, supervised, measured, scheduled.
Time and pacing
Each chapter is roughly 30–45 minutes of focused work plus an hour or two of tinkering; the project chapter that closes each phase runs 60–90 minutes. At a steady five-or-six-chapters-a-week pace, 71 chapters across 12 phases land around 10–14 weeks — a little longer than a tools course, because AI has more concepts that need to sink in.
Steady pace
Five or six chapters a week, one mini-project per phase.
Typical chapter
The project chapter closing each phase runs 60–90.
Access
Every phase project, before and after, kept for good.
One subscription. Every course.
AI Fundamentals Mastery isn't sold on its own. A single MJ Tech Institute plan unlocks it alongside every other course we've shipped — 9 today, with more added over time — at no extra cost.
Everything, included
Subscribe once and every course unlocks — this one and every future release. No per-course purchases, ever.
Cancel any time
Monthly and annual plans cancel from My account whenever you like. Your progress is always saved.
Start free first
The first 2 lessons of AI Fundamentals Mastery cost nothing, no card required, before you choose a plan.
Finish with 12 mini-projects and a working AI agent of your own
Phase 0 is free and takes about an hour. If it doesn't click, you've lost nothing.
Start the course today