Build and evaluate illegal-video classifier
Company: Google
Role: Data Scientist
Category: Machine Learning
Difficulty: hard
Interview Round: Technical Screen
Overview: This question evaluates competency in end-to-end Machine Learning system design, including multimodal modeling (vision, audio, text), data engineering for sparse, noisy, and imbalanced labels, robustness and abuse resistance, human-in-the-loop workflows, privacy/retention concerns, and operational metrics.
Read the full Google Data Scientist interview experience this question came from