Guest Speaker: Venu Satuluri
480 Dreese Labs
2015 Neil Avenue
Columbus, Ohio 43210
Machine Learning for Recommender Systems at Twitter
Recommender systems play a crucial role in Twitter's mission of helping people stay informed about the world, by surfacing the right tweets and trends and users that our users need to know about, at the right time. Twitter's uniqueness poses particular challenges for recommendations: content on Twitter evolves rapidly, and the relatively public nature of the Twitter platform further increases the space of candidates that can be recommended. In this talk, I will outline the architecture of our system that solves this and other challenges while supporting a mature product with a broad surface area. Then I will go into the details of how we have incorporated Machine Learning into our system and the practical difficulties we ran into in the process.
Bio: Venu Satuluri got his Ph.D. in Computer Science at the Ohio State University in 2012 with a dissertation focused on graph clustering and similarity search. Since then, he's been working at Twitter and currently leads the "MagicRecs" team at Twitter.
Host: Srini Parthasarathy