Three routes through the same material, depending on where you are starting from. Each step ends in something you can check, so you always know whether it worked.
Last reviewed
You have used a chatbot a few times and want to actually get good at this.
By the end you will have automated at least one recurring task, know why the model fails when it fails, and have shipped one small thing other people can use.
You write code and want to put a model into something real without it embarrassing you.
By the end you will have a feature backed by a model, returning validated structured output, grounded in your own data, with an eval set that tells you when a change made it worse.
You are not going to write code, and you still want to be genuinely good at this.
By the end you will have a small library of prompts you reuse, a clear sense of which tasks to trust a model with, and rules about what never goes near one.
| Path | Who it is for | You finish able to | Steps |
|---|---|---|---|
| Your first 30 days with AI | You have used a chatbot a few times and want to actually get good at this. | By the end you will have automated at least one recurring task, know why the model fails when it fails, and have shipped one small thing other people can use. | 4 |
| From API call to shipped AI feature | You write code and want to put a model into something real without it embarrassing you. | By the end you will have a feature backed by a model, returning validated structured output, grounded in your own data, with an eval set that tells you when a change made it worse. | 5 |
| Using AI well without writing code | You are not going to write code, and you still want to be genuinely good at this. | By the end you will have a small library of prompts you reuse, a clear sense of which tasks to trust a model with, and rules about what never goes near one. | 5 |
Pick by what you will do with it, not by how much you already know. Start with “Your first 30 days with AI” if you have used a chatbot and want to get properly good. Take the developer path if you write code and need a model in production. Take the no-code path if you are never going to write any.
No. One of these paths assumes you will never write a line of it and still takes you to genuinely capable. Prompting well, judging whether an output is right, and knowing where a tool helps and where it quietly does not are skills of judgement rather than syntax, and they are the ones most work actually needs.
Yes, and people do. The paths cover overlapping ground in a different order rather than teaching different subjects, so work you finish on one carries over. Nothing is locked, nothing expires, and there is no penalty for starting the developer path and deciding halfway that you wanted the other one.
Stuck on a word rather than a step?
The AI glossary explains every term these paths use, in plain English and without assuming you have met it before.