Go to our app
SCHOOLWHOOL CLASSROOM

Machine Learning
Detailed Course

Big ideas. Clear explanations. Real understanding.
A step-by-step path from curious beginner to building your own machine learning models.

Explore the course
Free for everyone No login required
2 modules A guided learning pathPython + ML Theory meets practiceYour own pace Room to pause and explore
YOUR NEXT LEARNING CHAPTER

A little theory. A lot of discovery.

4 free videos · More to come
MODULE 01 · Calculus needed for Machine Learning

Calculus & Derivatives

Build the calculus foundation needed for machine learning: understand derivatives visually and from limits, apply differentiation rules and the chain rule, find maxima and minima, and extend derivatives to gradient vectors. Learn why local and global minima differ and why iterative methods such as gradient descent are useful when solving directly is impractical.

11 chaptersWatch lesson MODULE 02 · Linear Regression

Geometric Interpretation of Linear Regression

Understand linear regression geometrically, from choosing a straight-line model and squaring its errors to training with gradient descent. Learn how to interpret coefficients, assess the model with R squared, recognise where a straight line breaks down, and connect the ideas to code. No machine learning or calculus background is required; reading graphs and multiplying is enough.

11 chaptersWatch lesson MODULE 02 · Linear Regression

Gradient Descent: how AI models and neural networks learn

Learn how gradient descent improves a model through repeated updates. Use derivative signs to choose a direction, apply the update rule by hand, build an iteration loop with a practical stopping criterion, and understand learning rates, shrinking steps, oscillation, divergence, and learning rate schedules. Extend the method to gradient vectors and problems without a closed-form solution.

8 chaptersWatch lesson MODULE 02 · Linear Regression

Gradient Descent for Linear Regression step by step

Derive the gradients of the linear regression loss and use them to train a model on a real dataset. Understand the cost of a full-data update, stochastic gradient descent and batch sampling, then implement SGD in Python and compare it with scikit-learn’s SGDRegressor. Explore why individual gradients disagree, why implementations differ slightly, and how learning rates and local minima affect training. Familiarity with graphs and derivatives is enough; the calculus lecture covers the required background.

5 chaptersWatch lesson
BUILT FOR THE CURIOUS

Understand the why.
Then build the how.

This course is designed to connect the ideas behind machine learning with the steps of building a model. Take it one concept at a time.

What we’ll explore

  • Turn raw data into useful features.
  • Understand how common ML algorithms learn.
  • Evaluate models and spot overfitting.
  • Bring the pieces together in a practical project.

A good place to start

Basic Python and school-level algebra will help. We’ll revisit the mathematical ideas as they come up.

No enrollment. No paywall.