Intermediate to senior

Machine Learning Interview Prep

Fifteen chapters from the learning problem and bias-variance to trees, neural networks, transformers, recommenders and ML system design, with tested NumPy code and diagrams.

15 chaptersUpdated 2026-10-02

Foundations

Chapter 1How ML Interviews Work and the Learning ProblemThe rounds you will meet, empirical risk minimisation, splits, baselines and how to answer a theory question.
Chapter 2Bias, Variance, Overfitting and ValidationThe decomposition, learning curves, cross-validation, hyperparameter search and leakage.
Chapter 3Probability, Statistics and A/B TestingBayes, distributions, MLE and MAP, hypothesis tests, power and experiment design.

Core models

Chapter 4Linear and Logistic RegressionNormal equation, ridge and lasso, log loss, softmax and from-scratch implementations.
Chapter 5Decision Trees, Random Forests and BoostingSplits and impurity, bagging, out-of-bag error, gradient boosting and tuning.
Chapter 6Classical Models and the Interview Cheat Sheetk-NN, naive Bayes, SVMs, model choice and short answers to the most-asked questions.
Chapter 7Unsupervised Learning: Clustering and PCAk-means, silhouette, DBSCAN, GMMs, PCA via SVD and anomaly detection.

Training and evaluation

Chapter 8Evaluation Metrics and Imbalanced DataConfusion matrix, ROC and PR curves, calibration, ranking metrics and cost-based thresholds.
Chapter 9Optimisation and RegularisationGradient descent, momentum, Adam, normalisation, L1 and L2, dropout and debugging training.
Chapter 10Feature Engineering, Data Quality and LeakageEncoding, scaling, missing values, feature construction and the kinds of leakage.

Deep learning and applications

Chapter 11Neural Networks and BackpropagationLayers, activations, backprop with a gradient check, vanishing gradients and architectures.
Chapter 12Sequence Models, Attention and TransformersRNN limits, scaled dot-product attention, transformer blocks, decoding and the KV cache.
Chapter 13Recommender Systems and RankingCollaborative filtering, matrix factorisation, the retrieval-ranking funnel, bias and cold start.
Chapter 14LLMs: Fine-Tuning, RAG and EvaluationTraining stages, prompt versus retrieve versus fine-tune, LoRA, evaluation, cost and security.
Chapter 15ML System Design and MLOpsA design framework, serving, deployment, drift monitoring, testing and responsible ML.
Header Logo