How to learn machine learning

Machine learning needs Python, some maths and many small experiments. Start with simple models and real datasets before deep learning.

Good for: People with basic programming who want to work with data and models. Typical time: 250–400 hours in total, depending on where you start.

The usual order of steps

1
Python for data≈ 40 h
NumPy, pandas and charts.
2
Maths you need≈ 50 h
Linear algebra, probability and statistics basics.
3
Classic models≈ 60 h
Regression, trees and how to test a model fairly.
4
Deep learning≈ 80 h
Neural networks with PyTorch or TensorFlow.
5
Projects≈ 60 h
Two projects on real data, shared publicly.

Free or paid?

Many good courses for each step are free; paid ones usually add a certificate, feedback from a teacher or a fixed schedule. Check the current price, dates and age limits on the provider’s own site before you pay.

Where to look for courses

Search these platforms for “Machine learning”. Compare ratings, prices and the date of the last update before you choose.

Get a plan with real courses

Stepora turns this outline into a plan with specific courses in the right order, matched to your level, budget and free time. It is free and takes half a minute.

Build my plan

The hours are rough, typical figures, not promises. Plans are made by AI and may contain mistakes.