Using ML to Power Live-stream Recommendations at a Global Scale

Friday, October 14, 2022 3 p.m. to 4 p.m.

Speaker: Omar Javed

From: Twitch (an Amazon company)


Twitch is a premier live social video platform with millions of unique streamers and tens of millions of daily viewers. In 2021, 1.3 trillion minutes of content was viewed on Twitch. Recommendations are a key feature: they allow viewers to discover and connect with streamers as well as find communities. In this talk, I will discuss the challenges of recommending live video content, and compare these to traditional Video on Demand recommendations. I will then present neural network based algorithms for generating live stream recommendations, highlighting how these algorithms solve the challenges associated with live content. I’ll go over the methods, offline experimentation, and evaluation metrics for live-streaming scenarios. Finally, I’ll talk about the solution’s success in a sitewide A/B test that significantly increased viewership.
I intend to kick-off the conversation with a discussion of my journey into ML based product development, and talk about key considerations for applied-ML research that powers products operating at a global scale.

For more info, please follow this link.

Read More


HEC: 101