
Eventbrite Tech Talk: Better Search at Scale + Discovery & Recommendations...
Event Information
Description
Join us for September's Eventbrite Tech Talk.
AGENDA
6:00 Doors Open, Dinner & Drinks
6:30 Welcome & Intro
6:45 Talk 1: Jonathan Dinu: Better Search at Scale: Leveraging Spark for Natural Language Processing + Q&A
7:00 Talk 2: John Berryman: Search & Recommendation Systems at Eventbrite + Q&A
8:00-8:30pm Mix & Mingle, play some pingpong or hit up the arcade
TALK 1 Better Search at Scale: Leveraging Spark for Natural Language Processing
Search, or more formally information retrieval, is fundamental to most people’s experience of the web. So much so that Google has become what is is today through its innovative use of algorithms to give better search results to users. And from a more abstract perspective, search allows us to turn the wealth of information on the internet into knowledge, or something we can actually use.
WHAT YOU'LL LEARN: In this talk I will walk through a practical example of how anyone with text data (which is most everyone) can leverage cutting-edge NLP techniques to improve search results. And you will see how to apply Spark’s machine learning APIs to build a topic model of project descriptions with LDA at scale.
ABOUT THE SPEAKER
Jonathan is currently the VP of Academic Excellence at Galvanize. Previously, he founded Zipfian Academy, which recently has been acquired by Galvanize. He first discovered his love of all things data while studying Computer Science and Physics at UC Berkeley. In a former life, he worked for Alpine Data Labs developing distributed machine learning algorithms for predictive analytics on Hadoop.
Jonathan has always had a passion for sharing the things he has learned in the most creative ways he can. At Galvanize, he gets to combine his two favorite things: humans and code. When he is not working with students you can find him blogging about data, visualization, and education at hopelessoptimism.com.
TALK 2 Search & Recommendation Systems at Eventbrite
On Eventbrite, users can serendipitously discovery events they will love. But making this possible this isn't easy. Events are short-lived and by the time we can build an adequate collaborative-filtering model, the event is already over. This talk discusses how we're overcoming technical challenges with a combination of collaborative-filtering and content-based methods.
WHAT YOU'LL LEARN: The nitty-gritty details of our problem space. Maybe it's similar to yours! Is a search engine a good technology for serving up recommendations? Our pain points and how we overcame them. What did we try? What failed? How did we scale? How did we determine the quality of our recommendations?
ABOUT THE SPEAKER John Berryman is a Senior Software Engineer on Eventbrite's Search and Discovery team. Working out of Eventbrite's Nashville office, John implements algorithms to provide personalized recommendations and search to users. John is also the author of Taming Search manning.com/turnbull/
NOTE: Bike parking is available on the 7th floor so come on up. The talks will be recorded and published on the Eventbrite engineering blog after the event.