Bias-Variance Tradeoff and Vanishing Gradients in Feedforward Networks

Company: Pinterest

Role: Machine Learning Engineer

Category: Machine Learning

Difficulty: hard

Interview Round: Technical Screen

Quick Answer: This question evaluates conceptual understanding of the bias-variance tradeoff and the vanishing-gradient problem in deep feedforward networks, two core machine learning fundamentals. It tests whether a candidate can reason about how model capacity affects generalization and why gradients shrink across layers during backpropagation, a common way interviewers probe theoretical depth alongside practical model-training intuition.