Data science interview prep
Practice the part of the interview where you have to think out loud.
Interactive notes on regression, probability, experiments, machine learning, SQL, and the judgment calls behind good answers.
Open regression 0 / 40 complete
Contents
The interview topics, in order
01–06Regression6/6 published
Choose a lesson to open its interactive notes. Published lessons are ready now; the remaining topics are mapped for a later pass.
- 01What a regression line is doingUnderstand regression as a prediction rule that minimizes total squared vertical error.
- 02Why least squares chooses that lineConnect residuals, the mean-only baseline, and the least-squares objective.
- 03SSE, R², coefficients, and uncertaintyEvaluate fit and interpret coefficient, confidence, and prediction uncertainty without confusing them.
- 04Holding other variables constantInterpret multiple regression, categorical variables, transformations, and interactions.
- 05When the line is lying to youDiagnose nonlinearity, heteroskedasticity, dependence, collinearity, and influential observations.
- 06Prediction, adjustment, and causalitySeparate predictive regression from causal claims and choose defensible controls.
07–11Probability5/5 published
Choose a lesson to open its interactive notes. Published lessons are ready now; the remaining topics are mapped for a later pass.
- 07Probability language and countingTranslate word problems into events and count outcomes without double-counting.
- 08Conditional probability and independenceDistinguish conditioning from independence and reason with probability trees and tables.
- 09Bayes’ theorem as belief updatingUpdate prior odds using evidence and explain base-rate effects.
- 10Random variables, expectation, and varianceCompute and interpret expectation, variance, covariance, and conditional expectation.
- 11Distributions and simulationRecognize interview-relevant distributions and use simulation to check probabilistic reasoning.
12–16Inference5/5 published
Choose a lesson to open its interactive notes. Published lessons are ready now; the remaining topics are mapped for a later pass.
- 12Sampling distributions and the CLTExplain sampling variability, standard error, and why many aggregate estimates become approximately normal.
- 13Estimators, bias, and maximum likelihoodCompare estimators by bias, variance, consistency, and likelihood-based reasoning.
- 14Confidence intervals, tests, and p-valuesInterpret intervals and p-values using repeated sampling, effect size, and practical significance.
- 15Power, sample size, and multiple testingConnect effect size and noise to power while controlling false discoveries and repeated looks.
- 16Resampling and Bayesian reasoningUse bootstrap and permutation logic and contrast frequentist intervals with Bayesian updating.
17–23Experimentation0/7 published
Choose a lesson to open its interactive notes. Published lessons are ready now; the remaining topics are mapped for a later pass.
- 17From product goals to metrics and estimandsTurn an ambiguous goal into a target population, causal estimand, primary metric, and guardrails.Coming soon
- 18Designing a trustworthy A/B testSpecify eligibility, randomization unit, exposure, duration, instrumentation, and an analysis plan.Coming soon
- 19Lift, uncertainty, and experiment decisionsEstimate effects, uncertainty, ratio metrics, and decision-relevant practical significance.Coming soon
- 20How experiments quietly failDiagnose sample-ratio mismatch, peeking, novelty, contamination, noncompliance, and logging failures.Coming soon
- 21Confounding, selection, and causal diagramsReason about confounders, mediators, colliders, selection, and identification using causal diagrams.Coming soon
- 22When randomization is unavailableExplain the assumptions behind difference-in-differences, regression discontinuity, matching, instrumental variables, and synthetic controls.Coming soon
- 23A complete marketplace experiment caseDesign and defend an experiment with interference, heterogeneous effects, guardrails, and a launch decision.Coming soon
24–32Machine learning0/9 published
Choose a lesson to open its interactive notes. Published lessons are ready now; the remaining topics are mapped for a later pass.
- 24Prediction targets, splits, and leakageDefine labels and prediction horizons and choose splits that match deployment reality.Coming soon
- 25Bias, variance, baselines, and validationDiagnose underfitting and overfitting and compare models against honest baselines.Coming soon
- 26Classification, thresholds, and calibrationInterpret logistic regression and select metrics and thresholds for imbalanced, cost-sensitive decisions.Coming soon
- 27Trees, random forests, and boostingExplain how tree ensembles learn, overfit, and trade interpretability for predictive power.Coming soon
- 28Regularization, features, and pipelinesUse ridge, lasso, transformations, and leakage-safe preprocessing pipelines.Coming soon
- 29Clustering, PCA, and anomaly detectionChoose and evaluate unsupervised methods without pretending unlabeled structure is ground truth.Coming soon
- 30Ranking, recommendations, and time-aware modelsRecognize the evaluation and validation patterns unique to ranking, recommenders, and temporal data.Coming soon
- 31Fairness, robustness, and model slicesInspect subgroup performance, distribution shift, interpretability, and asymmetric failure costs.Coming soon
- 32Serving, monitoring, drift, and AI evaluationReason about training-serving skew, latency, drift, retraining, rollback, and modern model evaluation.Coming soon
33–37Data practice0/5 published
Choose a lesson to open its interactive notes. Published lessons are ready now; the remaining topics are mapped for a later pass.
- 33SQL foundations, joins, and table grainWrite reliable aggregations and joins while preventing duplicate counts and NULL errors.Coming soon
- 34Windows, funnels, retention, and cohortsUse CTEs and window functions to solve event, funnel, retention, and sessionization problems.Coming soon
- 35Python, pandas, and analytical debuggingTransform and debug data clearly while explaining complexity, edge cases, and AI-generated-code errors.Coming soon
- 36EDA, data quality, and instrumentationAudit missingness, outliers, schema drift, event logging, leakage, and source-of-truth definitions.Coming soon
- 37Metric trees and root-cause diagnosisBuild metric trees and diagnose changes through segmentation, funnels, cohorts, and marketplace dynamics.Coming soon
38–40Interview practice0/3 published
Choose a lesson to open its interactive notes. Published lessons are ready now; the remaining topics are mapped for a later pass.
- 38Analytical storytelling and project deep divesPresent assumptions, evidence, tradeoffs, failures, and recommendations with executive clarity.Coming soon
- 39Your role-specific interview loopPractice the SQL/product, inference/analytics, applied modeling, or ML-systems loop appropriate to the target role.Coming soon
- 40A complete data science mock interviewComplete and diagnose a multi-round interview spanning data, statistics, decisions, and communication.Coming soon