Your recommender system (RecSys) can make you a lot of money, but behind the scenes maintaining or improving performance is a challenge. After talking with 50+ RecSys teams across e-commerce, media, and consumer marketplaces, Rubber Ducky Labs has seen the emergence of common problems as well as common solutions.
This talk will be a survey of design patterns for internal tools built by some of the world’s top RecSys teams, and attendees will learn about a number of initiatives they can drive internally to manage their RecSys operations.
Alexandra has a degree in computer science from Carnegie Mellon University. She started her career working on recommender systems in fashion tech at Polyvore (acquired by Yahoo in 2015), before joining an early pioneer in the ML tools space, SigOpt (acquired by Intel in 2020), where she led the Platform Team and earned two patents in ML tooling. As a co-founder of Rubber Ducky Labs, she’s building operational analytics for recommender systems teams.