Institute for Defense Analyses Data Scientist Interview Questions

Explain the bias-variance tradeoff mathematically and how L1 and L2 regularization change model complexity and weights.

Bias-Variance TradeoffRegularizationSupervised LearningInstitute for Defense Analyses

Tests your machine learning fundamentals and your ability to map model choice to threat-assessment use cases.

model trainingtechnical fundamentalsSupervised LearningInstitute for Defense Analyses

Tests your ability to evaluate causal impact with limited randomization and credible identification strategies.

experiment designCausal Inferenceprimary metricsInstitute for Defense Analyses

Explain your approach to model evaluation, including how you choose and interpret metrics for different ML problems.

PrecisionAccuracyRecallInstitute for Defense Analyses

Tests your product thinking and ability to align quantitative metrics with mission outcomes and risk controls.

Metricsuser valueProduct VisionInstitute for Defense Analyses

Tests your SQL proficiency for time-series comparisons and windowed aggregations.

Window FunctionsRankingRunning TotalsInstitute for Defense Analyses