Machine Learning curriculum

Start here — the syntax and mental models everything else builds on.

1. 1

   Foundations & generalization

   ~21 min · 3 lessons

   Frame any ML problem and explain why models over- or underfit.

   supervisedbias/variancecross-validationleakage

   • Learning paradigmsup next4m
   • Bias-variance tradeoff & overfitting12m
   • Validation strategies5m
2. 2

   Evaluation metrics & tradeoffs

   ~15 min · 3 lessons

   Pick and defend the right metric for the business cost of errors.

   precisionrecallROC/PR-AUCRMSENDCG

   • Classification metrics5m
   • ROC, PR curves & calibration5m
   • Regression & ranking metrics5m

Combine data from several places and reshape it into answers.

1. 1

   Classical algorithms

   ~15 min · 3 lessons

   Explain how classic models work and when to choose each.

   logisticL1/L2XGBoostk-means

   • Linear & logistic models5m
   • Tree-based ensembles5m
   • Clustering & distance5m
2. 2

   Feature engineering

   ~13 min · 3 lessons

   Prepare messy features without leaking the label.

   scalingone-hottarget encodingSMOTE

   • Scaling & transformations4m
   • Categorical encodings4m
   • Missing data & class imbalance5m

The techniques that show up in real interviews and real pipelines.

1. 1

   Production ML & modern concepts

   ~15 min · 3 lessons

   Discuss embeddings, deep nets and keeping models healthy after launch.

   embeddingsReLUdropoutdrift

   • Embeddings & vector search5m
   • Deep learning essentials5m
   • ML systems & monitoring5m