NVIDIA Data Scientist Interview Questions

Explain how you would implement a supervised ML model end to end, from preprocessing to validation and evaluation.

Cross-ValidationFeature EngineeringSupervised LearningNVIDIA

Approach for handling schema changes and data quality checks in a high-volume data lake pipeline.

schema evolutionData ModelingQualityNVIDIA

Define and calculate LTV to CAC for self-serve customers, then interpret what the ratio says about growth efficiency.

CACLTVfinancial metricsNVIDIA

Framework for identifying which user segment to target first for a new product use case.

User SegmentsUser NeedsUse CasesNVIDIA

Reason about power analysis when planning an experiment and choosing sample size.

ExperimentationPower AnalysisSample SizeNVIDIA

Choose the right classification metrics, and explain when precision, recall, and F1 score matter most.

F1 ScorePrecisionRecallNVIDIA