User Modeling and Recommendations at Pinterest

Pinterest is a catalog of ideas and a personalized recommendation engine. The data we’ve amassed over the years allows us to understand the tastes of people and ideas they’ll love but may not know exist. This way we can build human centered personalization engines that allow for magical experiences and adjust the types of Pins users see on the fly. The challenge is serving the right recommendation to the right person at the right time, from a pool of 100 billion objects in real-time. We are building some of the largest-scale recommender systems around that recommender systems that serve about 10 billion pins per day.




  • Charles Sugnet
  • Jerry Liu
  • Jure Leskovec
  • Mark Ulrich
  • Pong Eksombatchai
  • Pranav Jindal
  • Stephanie Rogers
  • Tao Cheng
  • Vanja Josifovski
  • Yuchen Liu