Part 8 of 9 · 1 min read
The map of the path
What each of the 15 levels teaches, in one plain line each.
The levels build on each other. Read them in order the first time.
| Level | Title | In one line |
|---|---|---|
| 0 | Engineering foundations | Ordinary software habits: Python, Git, tests, containers |
| 1 | How LLMs work | What a model is actually doing when it answers |
| 2 | LLM APIs and structured output | Calling a model from code and getting clean answers back |
| 3 | Prompting and workflows | Writing prompts and chaining fixed steps |
| 4 | Retrieval and RAG | Giving a model the right documents to read |
| 5 | Your first agent | Building the loop with tools and limits |
| 6 | Tools, MCP and computer use | Connecting agents to real systems safely |
| 7 | Memory and context | Deciding what the agent remembers and sees |
| 8 | Multi-agent systems | Several agents working together |
| 9 | Evaluation and observability | Measuring and tracing what agents do |
| 10 | Safety, security and governance | Stopping misuse and limiting damage |
| 11 | Production architecture | Running agents reliably at scale |
| 12 | Agent products and patterns | Coding, research, support and voice agents |
| 13 | Advanced topics | Fine-tuning, faster serving, small models |
| 14 | Capstone and career | A final project, debugging and interviews |
If you are brand new, follow the order. If you already build with models, you can start at Level 5 and return to earlier levels when a word is unfamiliar.