Kaggle Learn certificates

Learning plan

About 21 hours of focused study, split into 5 phases. Pick your weekly pace and the timeline adjusts.

Estimated finish: 3 weeks

What the exam covers

Source: official exam guide. Weights can change, so check before you start.

Before you start

• A free Kaggle account. Certificates are only issued to signed-in users who finish every exercise in a course
• No local setup needed, because everything runs in Kaggle Notebooks in the browser. Complete beginners should start with Intro to Programming

1. Week 1 · 4 h

   Python for data work

   • Write functions, loops, list comprehensions and dictionary operations without looking up the syntax
   • Import and use external libraries, and read a stack trace to fix a bug

   Use: Kaggle Learn – Python · Kaggle Learn – Pandas

2. Week 1 · 3 h

   Intro to Machine Learning

   • Train a DecisionTreeRegressor and a RandomForestRegressor with scikit-learn on the Iowa housing data
   • Split data with train_test_split and measure mean absolute error on validation data
   • Explain underfitting vs overfitting and tune max_leaf_nodes to find the sweet spot

   Hands-on: Make your first submission to the Housing Prices Competition for Kaggle Learn Users with your random forest, and note your public leaderboard MAE.

   Use: Kaggle Learn – Intro to Machine Learning

3. Weeks 1–2 · 4 h

   Intermediate Machine Learning

   • Handle missing values (drop, impute, impute with indicator columns) and categorical variables (ordinal vs one-hot encoding)
   • Build a scikit-learn Pipeline with a ColumnTransformer, then evaluate it with cross-validation
   • Train and tune XGBoost (n_estimators, early_stopping_rounds, learning_rate), and spot target leakage and train-test contamination

   Hands-on: Rebuild your housing model as a Pipeline with XGBoost, resubmit, and record how much your MAE improved over the random forest.

   Use: Kaggle Learn – Intermediate Machine Learning · Housing Prices Competition for Kaggle Learn Users

4. Week 2 · 7 h

   Feature engineering and deep learning

   • Rank features with mutual information, and create features with math transforms, counts, group statistics, k-means clusters and PCA
   • Apply target encoding safely without leaking the target
   • Build a Keras sequential network with dense layers, dropout and batch normalisation, and read learning curves to spot over- and underfitting

   Use: Kaggle Learn – Feature Engineering · Kaggle Learn – Intro to Deep Learning

5. Weeks 2–3 · 3 h

   Explainability and next steps

   • Compute permutation importance, partial dependence plots and SHAP values, and explain one prediction to a non-technical reader
   • Choose the next track (Computer Vision, Time Series, AI Ethics or the self-paced 5-Day AI Agents guide)

   Use: Kaggle Learn – Machine Learning Explainability · Kaggle Learn – Intro to AI Ethics · Kaggle 5-Day AI Agents Intensive (self-paced guide)

You're ready to book when…

• Each course you finished shows a certificate on your Kaggle profile
• You have a public notebook with a full pipeline (cleaning, encoding, XGBoost, cross-validation) that someone else can run
• You can explain a SHAP plot for one of your model's predictions

Exam tips

• Each course is a few hours of short lessons with an exercise notebook after each one. The certificate only unlocks when every exercise in the course is marked complete.