Machine Learning with Python

Build, train and judge models on data that is messier than a textbook's.

What the course covers

8 modules, 32 topics. Open any one to see what sits inside it.

1. 1.1 When machine learning is the wrong tool
2. 1.2 Supervised, unsupervised, and everything between
3. 1.3 Defining the target and the unit of prediction
4. 1.4 Choosing your metric before you build anything
1. 2.1 Loading, joining and inspecting a real dataset
2. 2.2 Missing values, outliers and leakage
3. 2.3 Encoding categories and scaling numbers
4. 2.4 Train, validation and test splits that do not lie
1. 3.1 Linear and logistic regression
2. 3.2 Decision trees
3. 3.3 Ensembles and gradient boosting
4. 3.4 Setting a baseline and beating it
1. 4.1 The confusion matrix and what it hides
2. 4.2 Precision, recall, and the cost of each kind of error
3. 4.3 Cross-validation
4. 4.4 Overfitting, underfitting and learning curves
1. 5.1 Feature engineering that earns its keep
2. 5.2 Selection and dimensionality reduction
3. 5.3 Handling imbalanced classes
4. 5.4 Explaining which features drove a prediction
1. 6.1 From a linear model to a network
2. 6.2 Training loops, loss and optimisers
3. 6.3 Regularisation and early stopping
4. 6.4 When a network is worth the cost
1. 7.1 Clustering
2. 7.2 Anomaly detection
3. 7.3 Working with text as features
4. 7.4 Using pre-trained models sensibly
1. 8.1 Packaging a model for use
2. 8.2 Monitoring drift after release
3. 8.3 Fairness, bias and harm
4. 8.4 Documenting a model someone else will inherit

By the end of this course you will be able to explain and apply:

Two hours, not two months Eight short modules. Finish in a day, or do one module a night for a week.

Depth beyond your coursework Pitched above taught content, for research, leadership and the work that follows the degree.

Built for a phone Watch on the bus, pause mid-module, pick up where you stopped. No download, no lab.

A certificate that can be checked Every certificate carries an ID your college or an employer can verify on our site.

About two hours in total — eight modules of roughly fifteen minutes. You can do it in one sitting or spread it across a week. Your place is saved either way.
No. There is no prerequisite course and no software to install. A phone, tablet or computer with an internet connection is enough.
Yes. This course is placed at Master's Degree for India, and the content is written for that level. If you are at a different stage, change your level at the top of the page and the courses will change with it.
A certificate for the course, carrying an ID that your college or an employer can check on our verification page. You keep access to the material afterwards.
Yes — ₹450 for this course on its own, or ₹1,450 for all five in this set, which works out cheaper if you want more than one.
There is a short quiz at the end of each module so you can check what stuck. It is not timed and you can retake it.
Change your country using the picker at the top. Level names, degree names and prices all switch to that country, and the courses adjust to its education system.

Wording shown for India. Pick a different country at the top and the terms, currency and programme names change with it.