How to Design a Metrics-First Recommender System

Katerina Zanos is a Principal Machine Learning Engineer at The Walt Disney Company, where she leads work on personalization for ESPN. Over the past decade, she’s shipped recommender systems and ranking models at The New York Times, Meta (Feed & Reels), and now Disney/ESPN, working end-to-end from product metrics and experimentation to large-scale ML infrastructure. With a background in both journalism and engineering, Katerina focuses on building recommendation systems for media products that are not only high-performing, but grounded in clear objectives and thoughtful product design.