Simulate and Compare Multi-Armed Bandit Strategies
Company: Wayfair
Role: Applied Scientist
Category: Machine Learning
Difficulty: medium
Interview Round: Onsite
Overview: Compare epsilon-greedy, UCB1, and Thompson sampling in a reproducible Bernoulli bandit simulation with reward, pseudo-regret, and repeated-run uncertainty.
Read the full Wayfair Applied Scientist interview experience this question came from