LFLearn with Faisal

World-Class AI Education for Rs. 0: A Free Four-Step Learning Roadmap

Faisal Adnan · 05 Nov 2026

Step 1: basics

  • Kaggle Learn (kaggle.com/learn): short, hands-on lessons in Python, pandas and intro machine learning. You write code from the first page.
  • Google Machine Learning Crash Course (developers.google.com/machine-learning/crash-course): core concepts such as linear regression, classification and loss, with interactive visuals.

Step 2: deep learning

  • fast.ai, Practical Deep Learning for Coders (course.fast.ai): free, top-down, for people with some coding experience who want to apply deep learning to practical problems.
  • Neural Networks: Zero to Hero (Andrej Karpathy, on GitHub and YouTube): you code and train neural networks from the basics, and build a GPT from scratch, line by line.

Step 3: large language models

  • Hugging Face Learn (huggingface.co/learn): courses on LLMs, agents and NLP, with free notebooks.
  • DeepLearning.AI short courses: one-hour courses on retrieval, agents and evaluation. Andrew Ng also teaches an "AI Prompting for Everyone" course there.
  • Generative AI for Beginners (Microsoft, on GitHub): 21 lessons with translations in many languages.

Step 4: build and ship

Pick one project, solve a local problem and publish it on GitHub. Not ten projects. One project that you can explain in five minutes and that someone else can run.

How to actually finish

  1. Set a weekly target of five hours, and put them in your calendar like a class.
  2. Move to the next step only when you have built something with the current one.
  3. Write a short note on each course: what you built, what broke, what you learned.
  4. Find one study partner. Finishing is easier with someone to report to.

Seedhi baat: nobody gets hired for watching courses. They get hired for something they built and can explain.

Sources

Part of the talk "The AI Era" by Faisal Adnan, Riphah International University, 6 November 2026.