◈AI Codex

Learning Paths

Which of these is you?

Pick the path that fits where you actually are — not a topic, not a product. Each one is sequenced to build on itself. Start at step 1, or jump to wherever you're stuck.

Haven’t done the product training yet? Anthropic’s Claude Academy is free and covers the products better than we do. Take it first — these paths pick up where it stops.

Which courses to take, by role →

Used AI before but new to Claude? Path 1 is the right starting point.

Start path 1 →

Never used AI before? Read this first — it takes 8 minutes.

Start here →
1

For individuals

Claude for your work

For anyone using Claude personally who wants to get noticeably better at it. What is safe to paste in, how to prompt well, what Claude is actually good at, the mistakes everyone makes first, and a workflow that survives past week three.

10 steps

~56 min

2

For managers & department heads

Rolling out Claude to your team

For the person responsible for getting their team actually using Claude. From picking the first task, to setting up a shared Project and connecting your tools, to knowing whether it's making a real difference and keeping it going past month three.

8 steps

~50 min

3

For IT & operations leads

Setting up Claude for your company

For the person asked to deploy Claude org-wide. Three stages — evaluation, deployment, ongoing management — with the exact decisions you need to make at each step: plan, policy, Projects, Skills, connectors, analytics and cost.

14 steps

~86 min

THE TWO ROLES DEFINING ENTERPRISE AI

For the person who owns AI internally

Becoming an AI Agent Manager

The operational playbook for the person responsible for running AI agents inside their own company — AI Agent Manager, AI Ops Manager, Agent Operator. First 90 days through ROI reporting, in the order you hit them, no developer background assumed.

8 steps

~67 min

For aspiring & working FDEs

Becoming a Forward Deployed Engineer

The role, the market, the path in, and Anthropic's own FDE training — then the technical playbook of what FDEs actually build, and the three parts of the engagement nobody writes down: scoping it, surviving the client's data, and handing it over so it still runs in month six.

12 steps

~115 min

WHEN YOU'RE READY TO BUILD

For founders

Building your first AI product

For founders building a product on Claude: choosing the problem, validating it, picking a simple stack, the mistakes that kill early AI products, shipping something that lasts, and pitching it.

10 steps

~64 min

FOR CLAUDE CODE USERS

For teams using Claude Code

Setting up Claude Code for your team

Six guides on the .claude folder: what to configure, what to write in CLAUDE.md, which decisions to make as a team, how to keep it accurate, and what the agent layer looks like when you are ready for it.

6 steps

~44 min

FOR ENGINEERING LEADS & OPS TEAMS

For engineering leads & ops teams

Building your internal AI stack

Eight guides on the architecture that makes AI agents actually work at scale: a single internal MCP server, the right data access patterns, caching that eliminates cold start, access control that works, and the skills layer that turns repeated workflows into one-command automation.

8 steps

~74 min

FOR DEVELOPERS

{ }

// technical

Building with the Claude API

Implementation guides that assume you can code. The messages array, streaming, RAG, evals, tool use, prompt caching, cost optimization, persistent memory, auth, and rate limiting.

21 steps

~162 min