Handle p≈n linear regression with L1
Company: Google
Role: Data Scientist
Category: Machine Learning
Difficulty: medium
Interview Round: Technical Screen
Overview: Fit a linear regression with 500 predictors and only 600 observations, explaining why ordinary least squares breaks down and how L1 regularization helps. Cover correlated predictors, tuning and evaluation, comparison with ridge and elastic net, and what selection means for inference on coefficients.
Read the full Google Data Scientist interview experience this question came from