Presented By
O’Reilly + Intel AI
Put AI to Work
April 15-18, 2019
New York, NY

Unifying Twitter Around a Single ML Platform

Yi Zhuang (Twitter), Nicholas Leonard (Twitter)
2:40pm3:20pm Wednesday, April 17, 2019
Implementing AI
Location: Rendezvous
Secondary topics:  Deep Learning and Machine Learning tools, Media, Marketing, Advertising, Platforms and infrastructure

Who is this presentation for?

ML Practitioners / ML engineers in large organizations



Prerequisite knowledge

* Basic understanding of ML. * Basic understanding of computer engineering. * (Nice-to-have) Beginner knowledge in TensorFlow.

What you'll learn

* What does a scalable production ML pipeline look like at Twitter today? * Lessons we learned while unifying Twitter's ML tech stack. * Which problems do ML solve at Twitter? * How to create platform that is flexible enough to serve a diversity of ML needs?


Twitter is a 4000+ employee company with many ML use cases. Historically, there are many different ways to productionize ML at Twitter. Across the company, we had users of TensorFlow, Lua Torch, PyTorch, Scikit Learn, VW, XGBoost, and several other Twitter in-house solutions. In this session, we are going to discuss:

  • the pain points of having fragmented ML practices;
  • how the Twitter Cortex team is unifying Twitter teams around a single ML platform;
  • current status of the ML platform and who are the internal customers;
  • how the production pipeline looks like and how customers use it;
  • where is the unified ML production platform heading in the future.
Photo of Yi Zhuang

Yi Zhuang


Yi Zhuang is a software engineer at Twitter, where he tech leads a group of people to build platform and infrastructure for working with ML models. Currently, he works on unifying Twitter around a single ML framework & infrastructure, bringing together users of Lua Torch, PyTorch, Scikit Learn, XGBoost, VW, and other in house ML solutions. Previously, Yi led a group of people to develop a trillion-document scale distributed search engine at Twitter. Yi holds an MS in computer science from Carnegie Mellon University. He loves cats and enjoys pondering over all things technical and logical.

Photo of Nicholas Leonard

Nicholas Leonard


Nicholas Leonard graduated from the Royal Military College of Canada in 2008. He obtained an MS in computer science from University of Montreal in 2014. Nicholas is a software engineer at Twitter Cortex. He was a core contributor to Lua Torch, and currently works with TensorFlow as part of the DeepBird team.

Leave a Comment or Question

Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?

Join the conversation here (requires login)