metacareerguide.com
MCG Classroom

Learn what AI actually is.

Two courses and two exam study desks. Free, self-paced, and written for adults — no technical background needed to start. One free account opens all of it.

Create your free learner account Read module 01 free — no account

Built in the open. Course 01 modules 01–03 and Course 02 modules 01–02 are live and free with an account — more arrive regularly.

Two courses. Neither needs the other.

Every module released so far is free with a Metacareerguide account. Later modules in each course will sit behind a paid tier when it opens — everything currently published stays free.

Course 01 · Beginner

Understanding & Using LLMs

How large language models learn, how to prompt them well, and how to tell when they’re wrong. From tokens to retrieval to evaluation.

16 modules · self-paced →
Course 02 · Intermediate

Agentic AI & the Future of Work

Systems that plan, use tools, and take real actions — how they’re built, where they fail, and who’s accountable when they do.

16 modules · 01–02 open →

Two certification study desks.

Full practice benches for two industry exams. Both free with a Metacareerguide account, both scored the way the real exam is, and both with a drill that only asks what you got wrong.

Anthropic · CCAR-F

Claude Certified Architect

Agentic architecture, Claude Code, prompt engineering, tool design and context management. 128 scenario questions and a timed practice exam.

Open the study desk →
EC-Council · CEH 312-50

Certified Ethical Hacker

Updated for v13. All twenty official modules across seven study domains, including AI-augmented attack and defence. A full four-hour timed exam at the real length.

Open the study desk →

Independent study aids. Not affiliated with, endorsed by or accredited by Anthropic or EC-Council. CEH requires an eligibility application or official EC-Council training before you can sit it.

Certification prep · free

Claude Certified Architect — Foundations

128 scenario questions, a full timed practice exam scored to the real 720 pass line, and a drill that only asks what you got wrong. Free with your learner account.

Open the Study Desk

Independent study aid. Not affiliated with, endorsed by, or certified by Anthropic. The $125 exam fee goes to Anthropic; nothing goes to us.

ITEM 24 / 60 · 78:14 left
A billing agent returns half-written JSON. Logs show those turns ended with stop_reason "max_tokens".
What is the correct handling in the loop?
Parse what arrived and default the missing fields
Treat the turn as incomplete — never ship the fragment
Inside a module

Figures you step through, not diagrams you squint at.

Abstract ideas — how a model reads a sentence, how agents hand work to each other, where a request actually goes — are built up one stage at a time, at your reading pace. Nothing animates until you ask it to.

Every module pairs the explanation with a hands-on exercise using free tools, then three questions that tell you honestly why the tempting wrong answer is tempting.

See it in module 01 — no account
Figure 1 — watch a sentence become tokens
The791 ·cat8415 ·sat7731 ·on402
Next stage STAGE 1 OF 3

Each block is one token. The small number beneath is its ID — the actual value the model receives.

Free learner account

One account. Everything open.

Your account is what remembers where you are — so you can study for twenty minutes on a Tuesday, close the tab, and pick up exactly where you left off.

Your progress follows you

Complete a module on your laptop, carry on from your phone. Nothing to keep track of yourself.

Both study desks unlock

CCAR-F and CEH, with every practice question and both timed exams. No upgrade prompt in the way.

New modules arrive to you

Opt into the weekly note and you’ll know when the next module goes live, along with what’s worth reading meanwhile.

Create your free account

An email address and a password — that’s the whole form. No card, no cost, and the newsletter is a checkbox you can untick.