Learn Machine Learning

Anime-inspired roadmaps, projects, and resources to build your machine learning foundations and portfolio.

Start your Machine Learning Journey

Beginner Arc: Learn Python, Math, and ML Fundamentals

Model Arc: Train models, evaluate performance, build projects

Portfolio Arc: Build 2-3 projects, deploy, get interview ready

How to learn machine learning in five steps

Hey friends! Start with a small task you can finish, then build on it. This path takes you from your first Python functions to a project you can explain. Move at your own pace and practise as you go.

1. Get comfortable with Python

Learn variables, loops and functions, then use pandas to explore a small table. Your first checkpoint is loading a CSV and explaining what each column means. Pick a starting point from our Python courses.

2. Learn the maths behind your results

Start with averages, variation and probability, then learn how vectors represent data. Practise by comparing the mean and median of a dataset. Our statistics and probability guide helps you choose what to study next.

3. Train your first model

Choose a small regression or classification problem. Separate training data from test data, fit a simple model and compare its predictions with a basic baseline. Follow one of our introductory machine learning courses as you build.

4. Check where your model fails

Inspect mistakes and choose a metric that matches the problem. For a classifier, count false alarms and missed cases. Work through our model evaluation guide and interactive notebooks to practise making those decisions.

5. Finish a project you can explain

Choose a question you care about from our dataset collection. Document the data, baseline, evaluation and limitations. Your finished project should let another learner reproduce your result and understand one thing you would improve.

Try it today: change the model’s alert threshold

Ready for a practical challenge? In the classification threshold notebook in Colab, follow the setup cells, then compare thresholds of 0.3, 0.5 and 0.7 using the same predictions. Before running the comparison, write down which threshold you expect to produce the most alerts.

Record the false positives and false negatives at each threshold. Choose a threshold for a situation where missing a positive case is costly, then explain the extra false alarms you would accept. If you are new to Python, begin with step 1 and return to this challenge after training your first classifier.

Keep learning with the free 90-day roadmap

Get the free PDF to organise these steps into a 90-day plan, with weekly tips and learning resources by email.

Machine Learning is a field with many different resources. Explore them here:

machine learning books

Books: Learn from the best ML, Python, and AI books

Explore →

 machine learning courses

Courses: Curated Learning paths from top platforms

Explore →

machine learning tutorials

Tutorials: Hands-on guides for real ML skills.

Explore →

Datasets for machine learning

Datasets: Practice with real-world datasets.

Explore →

machine learning blogs

Blogs: Top Curated ML blogs and insights.

Explore →

machine learning code snippets

Code Snippets: Useful code examples and snippets.

Explore →

Get the ML Starter Kit

Everything you need to learn faster: full roadmap, project templates, datasets, and portfolio.

The Machine Learning Starter Kit

A 172-page practical handbook: 26 chapters from your first line of Python to a deployed model. Every code listing was actually executed — the outputs you read are the real ones.

See what’s inside →

Instant PDF download · yours forever


Also, find some cool articles about Machine Learning and Data Science with our Medium James Thorn.