Put AI to work
June 26-27, 2017: Training
June 27-29, 2017: Tutorials & Conference
New York, NY

How AI powers the Comcast X1 voice interface

Jan Neumann (Comcast), Ferhan Ture (Comcast), Shahin Sefati (Comcast)
4:00pm4:40pm Thursday, June 29, 2017
Implementing AI
Location: Gramercy East/West Level: Intermediate
Secondary topics:  Deep Learning, Media, Natural Language, Speech and Voice, User interface and experience

Prerequisite Knowledge

  • General familiarity with NLP and deep learning

What you'll learn

  • Explore the technology behind Comcast's X1 voice remote and lessons learned deploying its AI solution to millions of customers


Comcast is the largest cable provider in the US, reaching more than 30 million customers. Recently, Comcast reinvented the way its customers can access content and other services. With the X1 voice remote, customers can interact with the X1 entertainment system and other Comcast services such as billing or Xfinity Home services via spoken language.

Jan Neumann, Ferhan Ture, and Shahin Sefati explain how AI enables Comcast to understand what you are looking for when you talk to the X1 voice remote and how Comcast scaled the voice interface to answer millions of voice queries every single night. Jan, Ferhan, and Shahin share how Comcast uses natural language processing (NLP) and deep learning to identify the intent behind a customer’s query across multiple domains and answer even general questions, such as “How old is Tom Hanks?” Along the way, they cover the lessons learned going from a research project to production and scaling the system to answer millions of queries per day.

Topics include:

  • How RNNs have become the centerpiece of the NLP framework
  • A novel solution to the slow speed of RNNs
  • Comcast’s NLP pipeline to control the trade-off between efficiency and accuracy
  • Training and deploying models using NVIDIA Docker
  • How to ensure products follow the correct logic even under high levels of ambiguity
Photo of Jan Neumann

Jan Neumann


Jan Neumann leads Comcast’s Applied Artificial Intelligence Research Group, which combines large-scale machine learning, deep learning, NLP, and computer vision to develop novel algorithms and product concepts such as voice interfaces, virtual assistants, and video and IoT analytics that improve the experience of Comcast’s customers. Previously, Jan worked for Siemens Corporate Research on various computer vision-related projects, such as driver assistance systems and video surveillance. He has published over 20 papers in scientific conferences and journals and is a frequent speaker on machine learning and data science. He holds a PhD in computer science from the University of Maryland, College Park.

Photo of Ferhan Ture

Ferhan Ture


Ferhan Ture is a member of the Comcast Labs DC research group, where he focuses on combining deep learning and natural language processing to build data-driven solutions for various Comcast products involving language, including the algorithms behind the voice-enabled remote controller for the X1 entertainment system. He is a frequent speaker at conferences worldwide.

Photo of Shahin Sefati

Shahin Sefati


Shahin Sefati is a senior researcher in the Applied AI research group at Comcast in Washington, DC, where he leads the personalization research for search, browse, recommendation, and content discovery. Shahin’s research interests include machine learning and big data analytics.