← Learn
Learn · AI

Agents and workflows

Seven lessons on building AI systems that take several steps: the agent loop and its tools, the common workflow patterns, guardrails, retries and failure handling, keeping a person in the loop, pulling structured data out of invoices, statements and scanned PDFs, and routing each request to the right model. General patterns, not one framework.

  1. The agent loop and its tools

    What makes something an agent, the loop at its heart, the limits every loop needs, and when not to build one.

    6 questions

  2. Orchestrating multistep workflows

    Chaining, routing, running in parallel, orchestrator and workers, and check-and-revise loops.

    6 questions

  3. Guardrails

    Checks on what goes in, what comes out and what the system is allowed to do, plus budgets and logs.

    6 questions

  4. Retries and failure handling

    Which failures to retry, backing off with jitter, idempotency keys, retrying with feedback, and where failures end up.

    6 questions

  5. Keeping a human in the loop

    When a person approves, when they review, what they need to see, and how their decisions improve the system.

    6 questions

  6. Extracting data from documents

    Invoices, statements and scanned PDFs: getting the text, pulling out the fields, checking them and loading them.

    6 questions

  7. Routing between models

    Sending each request to the cheapest model that can handle it: rules, classifiers, cascades and fallbacks.

    6 questions
More info

How it works

Each lesson is a short read with examples, then a few practice questions. Some ask you to type an answer (a number, an address, what a piece of code prints), some have you pick one, and some have you tap steps or lines into the right order. You can try again as often as you like; after two misses there’s a button to show the answer.

What do I need to know first?

How LLMs work first, especially the lessons on JSON output and tool calling. The examples are short Python you only need to read.

The other AI and data engineering courses

Where is my progress saved? In this browser, like the rest of the site. Nothing to sign up for.

“Python” is a registered trademark of the Python Software Foundation. This course isn’t made by or affiliated with it.