Machine Learning Foundations

Tracks

1. 01🧠What Machine Learning Is

   0/10 lessons

   Problems, targets, labels, and costs

   ML starts when labeled examples can teach a reusable mapping and the cost of being wrong can be measured. Frame the problem before touching an estimator.

   Lesson list (10)

   • 01
     What Machine Learning Actually Is~32 min · foundation, machine-learning, framing
   • 02
     Rules, Heuristics, Statistics, and ML~28 min · framing, rules, statistics
   • 03
     ML Task Families~26 min · task-types, framing, supervised
   • 04
     Supervised, Unsupervised, and Beyond~26 min · supervised, unsupervised, self-supervised
   • 05
     Features, Labels, Examples, Datasets~30 min · features, labels, tabular-data
   • 06
     Prediction, Explanation, and Decision-Making~26 min · framing, decisions, interpretability
   • 07
     Good and Bad ML Use Cases~24 min · framing, scoping
   • 08
     The Hidden Cost of ML~28 min · mlops, lifecycle
   • 09
     Classical ML, Deep Learning, LLMs, and RAG~30 min · framing, deep-learning, llm
   • 10
     Roadmap: What You Will Be Able to Do~22 min · roadmap, metacognition

   ✓Quiz · 4 questions→

2. 02🗂️Data First

   0/8 lessons

   Leakage, splits, missingness, preprocessing

   Most production wins live in data quality, leakage prevention, honest splits, and preprocessing pipelines.

   Lesson list (8)

   • 01
     Better Data Beats Fancier Models~28 min · data-centric, labeling, framing
   • 02
     Tabular Data Intuition~26 min · tabular, eda, pandas
   • 03
     Target Leakage~32 min · leakage, splits, evaluation
   • 04
     Train, Validation, and Test Split~30 min · splits, validation, evaluation
   • 05
     Class Imbalance~28 min · imbalance, metrics, sampling
   • 06
     Missing Values and Outliers~30 min · missing, outliers, preprocessing
   • 07
     Feature Scaling and Encoding~28 min · preprocessing, scaling, encoding
   • 08
     Data Pipelines with scikit-learn~30 min · pipelines, sklearn, reproducibility

   ✓Quiz · 4 questions→

3. 03📈Regression

   0/5 lessons

   Continuous targets, residuals, regularization

   Predicting numbers honestly: pick the metric that matches the cost, regularize linear models, and know when linear breaks.

   Lesson list (5)

   • 01
     What Regression Means~24 min · regression, framing
   • 02
     Linear Regression Intuition~28 min · linear, regression
   • 03
     Regression Metrics: MAE, RMSE, R²~26 min · metrics, regression
   • 04
     Underfitting, Overfitting, and Regularization~30 min · regularization, ridge, lasso
   • 05
     When Linear Works and When It Breaks~26 min · linear, limitations

   ✓Quiz · 4 questions→

4. 04🎯Classification

   0/4 lessons

   Probabilities, thresholds, and costs

   Treat probabilities as the model output. The threshold is a business decision that belongs in the design doc, not the training script.

   Lesson list (4)

   • 01
     Classification Basics~26 min · classification, logistic-regression
   • 02
     Confusion Matrix and Metrics~28 min · metrics, confusion-matrix
   • 03
     ROC-AUC vs PR-AUC~28 min · metrics, roc, pr-auc
   • 04
     Threshold Tuning and Business Cost~30 min · threshold, cost-matrix

   ✓Quiz · 4 questions→

5. 05🌲Model Selection, Trees, and Boosting

   0/7 lessons

   Cross-validation, capacity, tree ensembles

   Cross-validate, tune, diagnose bias vs variance, and graduate from a single tree to random forests and gradient boosting.

   Lesson list (7)

   • 01
     Cross-Validation~28 min · cv, validation, model-selection
   • 02
     Hyperparameters~28 min · hyperparameters, tuning, search
   • 03
     Overfitting and Underfitting~26 min · overfitting, underfitting, diagnostics
   • 04
     Bias-Variance Tradeoff~26 min · bias, variance, theory
   • 05
     Decision Trees~28 min · trees, interpretable
   • 06
     Random Forests~28 min · random-forest, bagging, ensembles
   • 07
     Gradient Boosting and Tabular Champions~32 min · gradient-boosting, lightgbm, xgboost

   ✓Quiz · 4 questions→

6. 06🌌Unsupervised Learning

   0/4 lessons

   Structure without labels

   Cluster, reduce dimensions, detect anomalies, and turn unstructured inputs into embeddings ready for classical models.

   Lesson list (4)

   • 01
     Clustering~28 min · clustering, kmeans, dbscan
   • 02
     Dimensionality Reduction~28 min · pca, umap, tsne
   • 03
     Anomaly Detection~28 min · anomaly, outliers, isolation-forest
   • 04
     Embeddings as Features~26 min · embeddings, representation

   ✓Quiz · 4 questions→

7. 07🔬Interpretation and Monitoring

   0/5 lessons

   Errors, importance, calibration, drift

   Read the errors, audit importance, calibrate probabilities, monitor drift, and respect the limits of correlation.

   Lesson list (5)

   • 01
     Error Analysis~28 min · errors, diagnostics
   • 02
     Feature Importance~28 min · importance, interpretation
   • 03
     Calibration~28 min · calibration, probabilities
   • 04
     Drift Monitoring~28 min · drift, monitoring, production
   • 05
     Causality Limits~24 min · causality, limits

   ✓Quiz · 4 questions→

8. 08🛠️Practical ML Workflow

   0/5 lessons

   From notebook to artifact

   Wrap the whole loop into a reproducible pipeline, ship it as one artifact, deploy it lightly, and decide when ML is even the right tool.

   Lesson list (5)

   • 01
     End-to-End Workflow~28 min · workflow, end-to-end
   • 02
     Feature Engineering as Leverage~30 min · features, engineering
   • 03
     Pipeline Artifact~26 min · artifact, serialization, deployment
   • 04
     Lightweight Deployment~30 min · deployment, fastapi
   • 05
     Rules vs ML vs DL vs LLMs~28 min · framing, decisions

   ✓Quiz · 4 questions→