A map of topics vs. a path you can follow
A roadmap answers 'what are the topics, roughly in what order'. That's a real question, and a diagram answers it well. It leaves three harder ones unanswered: which resource should I use for this topic, is it current and good, and what do I do after I finish it? Answering those for every box is hours of searching, comparing and second-guessing. The moment one link is outdated or one step is ambiguous, momentum dies.
An ordered path answers all three. path·ai's curated tracks (computer vision, nlp and transformers, generative ai, reinforcement learning) break each topic into modules, and each module already points at a resource worth your time, the code to practise it, and a checkpoint that tells you whether to move on or revisit. You spend your time learning instead of curating your own syllabus.
Where the resources come from, and what path·ai isn't
path·ai doesn't host courses or replace them. It links the good ones. The curated tracks point at established, mostly free material (think foundational university courses, canonical papers, respected tutorials), and the generated paths pull current resources with live web search so the links aren't years stale. The depth still comes from those underlying resources. path·ai is the ordering, the code and the checkpoints around them.
The limits are worth stating. path·ai isn't an accredited course and gives no certificate, it's no job guarantee, and there's no community or mentor attached. The checkpoints are self-checks, not graded exams. What you get is the part a static roadmap leaves out: a concrete, ordered route with the resources and code already attached, for free and with no account.
A static roadmap vs. path·ai
| Static roadmap (diagram / list) | path·ai | |
|---|---|---|
| What it gives you | A map of topics and order | Ordered modules you follow step by step |
| The resources | You find and vet them | Linked for each step |
| Runnable code | Not included | Paired with every module |
| Knowing you're ready | On you | A checkpoint per module |
| Staying current | Often goes stale | Generated paths use live web search |
| Price | Usually free | Free, no account |
frequently asked
What is the best way to follow a machine learning roadmap?
Pick a roadmap whose steps link the actual resource and some code, not just the topic names, then do them in order with a way to check you understood each one. That's what path·ai is: ordered modules, a linked resource and runnable code per step, and a short checkpoint.
Does path·ai include the learning resources or just the topics?
The resources. Each module links a specific resource to read or watch plus runnable code, rather than only naming the topic and leaving you to search.
Do I need a degree to follow it?
No. path·ai is a free, self-serve path to public resources. It isn't accredited and gives no certificate. It orders the route and links what to use at each step.
Is it really free?
Yes. path·ai is free with no account and no paywall between steps.
Last updated June 7, 2026