Building a Recommendation Engine at ESPN Without Slowing Delivery: AI as Your Team's Multiplier

How do you modernize a recommendation system while the product keeps shipping? In this talk, Katerina will share how the team at ESPN evolved the app's personalization stack from heuristics to an ML-first recommender system, where each stage delivered value on its own while building toward the next.
She'll show where AI tooling acted as a force multiplier, compressing iteration cycles across documentation, evaluation, and debugging, and helping them spot impact, so they could focus on the right levers.
You'll leave with key takeaways on incremental modernization and where AI can truly accelerate your team's velocity.