Loading…
Back To Schedule
Saturday, October 20 • 12:00pm - 12:30pm
Production Ready Real-Time Recommendations in About a Month

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Feedback form is now closed.
At Joymode it took a machine learning data scientist and a fraction of a cloud devops engineer about a month to build the leanest customer facing, real-time personalization engine we could devise. In this presentation I will describe the end-to-end solution, which includes Docker for environment management, Jupyter and Papermill for batch training matrix factorization models, and Flask for delivering the recommendations via a RESTful web service. I'll also cover the strategic bets we are making, like dynamic prototyping, long term AB testing, and extremely rapid delivery of AI solutions for in-flight iteration.


Speakers
avatar for William High

William High

Head of Data Science, Joymode
Will is Head of Data Science at Joymode in Los Angeles. He was previously a data scientist at Netflix LA and occasionally serves as a pro bono Data Ambassador for DataKind. Will received a PhD in astrophysics from Harvard and a BA from Berkeley and has held research posts at Caltech... Read More →


Saturday October 20, 2018 12:00pm - 12:30pm PDT
Ballroom # 408A