You don’t need to pay for a bootcamp to understand AI. This path uses free, high-quality resources and moves from intuition to hands-on practice.

Stage 1 — Intuition (1–2 weeks)

  • 3Blue1Brown’s neural networks series. Beautifully animated videos that explain what a neural network is, how it learns and how transformers work, with almost no prerequisites.
  • Our Beginner’s Corner. Short explainers on machine learning and generative AI.

Goal: explain to a friend what “training a model” means.

Stage 2 — Machine learning foundations (3–4 weeks)

  • Google’s Machine Learning Crash Course. Covers regression, classification, overfitting and evaluation with interactive exercises.

Goal: understand training data, test data and why models fail on new situations.

Stage 3 — Deep learning by doing (6–8 weeks)

  • fast.ai, Practical Deep Learning for Coders. A top-down course: you train working models from the first lesson, then learn the theory behind them. Requires basic Python.

Goal: train and evaluate an image or text model yourself.

Stage 4 — Language models and modern tools (4 weeks)

  • Hugging Face Learn. Free courses on transformers, LLMs, agents, diffusion models and more, using the open-source libraries most practitioners use.

Goal: fine-tune a small model and build a simple application on top of an open model.

Tips

  • Code along. Watching is not learning.
  • Build one small project per stage and write a short note about it.
  • Don’t get stuck on math at the start; return to it when a concept needs it (see our Math for AI guides).
  • Read primary sources once comfortable: papers explained in our Research section are a good bridge.

How to stay motivated

  • Set a weekly schedule — for example three sessions of one hour — rather than vague goals.
  • Learn with others: study groups, online communities or a friend following the same path.
  • Build small things early. Even a simple classifier or a chatbot using an API makes concepts real.
  • Keep a learning log of what you understood and what confused you.
  • Accept confusion. Feeling lost is normal; revisit difficult topics later with fresh eyes.

Adapting the path to your goals

  • If you want to use AI at work: focus on Stage 1 and Stage 4, plus prompting and evaluation skills.
  • If you want to build AI applications: emphasize programming, Stage 4 and the developer learning path.
  • If you want to become an ML engineer or researcher: go deep on Stages 2 and 3, with more mathematics.
  • If you work in policy or management: prioritize intuition, limitations, risks and regulation.

A weekly routine that works

A sustainable routine for someone with a full-time job:

  • Two weekday sessions (45–60 minutes): watch or read one lesson and take notes in your own words.
  • One weekend session (2 hours): practice — exercises, a notebook, or a small project.
  • Ten minutes daily (optional): review flashcards of key concepts or read one short article.

Every four weeks, pause to review: what can you now explain to a friend? What’s still unclear? Adjust the plan rather than abandoning it when life gets busy.

Signs you are ready for the next stage

  • You can explain the previous stage’s key ideas in your own words.
  • You completed at least one exercise or small project without copying a solution.
  • You know where to look when you get stuck.

If not, spend another week practicing — consolidation makes later stages much easier and prevents the frustrating feeling of not understanding anything in advanced material.

What to watch next

  • New free courses on agents and reasoning models.
  • AI tutors that personalize learning paths.
  • Updated university course materials published online.
  • Certifications tied to AI skills and regulation.

Key takeaways

  • You can learn AI for free through a staged path from intuition to modern tools.
  • Build intuition first, then foundations, deep learning and language models.
  • Hands-on projects make concepts stick.
  • Adapt the depth of each stage to your goals.

Frequently asked questions

How long does this learning path take?

Roughly four to five months at a few hours per week, depending on your background and depth.

Do I need to know programming?

For the intuition stage, no. For later stages, basic Python is very helpful.

Are free resources as good as paid ones?

Many free courses from universities and well-known educators are excellent; paid options mainly add structure, support or certificates.