Search at GIPHY: Staying on Top of the Zeitgeist

Hosted by Analytics & Data Science by Dataiku NY

Public group


Hi search engine enthusiasts,

We're excited to provide you with one of the brightest and trendiest examples in which search engines can make or break a product. Yael Elmatad, Lead Data Scientist at GIPHY, will explain how she and her team develop their search engine to remain at the top of the social media frenzy.

Prior to this, Patrick Masi-Phelps, Data Scientist at Dataiku, will walk through image classification using deep neural nets in Dataiku, improving models using transfer learning, and applying similar techniques to the classification of GIFs.

Search at GIPHY: Staying on Top of the Zeitgeist

While on the outside GIPHY search may seem like a standard search problem, being on the top of the zeitgeist is an increasingly viral world is a constant challenge. GIPHY has devised ways to make sure their users are getting the most relevant search via models that allow them to exploit known performant content while experimenting with potentially successful content. This model ensures relevant content is surfaced to the end user while still trying to explore the space to ensure that new and exciting content has a chance to propagate up towards the top of the page.

Speakers' bios:

Yael Elmatad is the Lead of the Search & Discovery Division at GIPHY. Prior to that, she spent 4 years at Tapad (a Marketing Technology company based in NYC, acquired by Telenor in 2016) as a Senior Data Scientist. Yael earned her undergraduate degree in Chemistry from NYU in 2006 and graduated as Valedictorian of the College of Arts and Sciences. She earned her Ph.D. from UC Berkeley in Theoretical Physical Chemistry in 2011.

Patrick Masi-Phelps is a Data Scientist at Dataiku, where he helps clients build and deploy predictive models. Before joining Dataiku, he studied math and economics from Wesleyan University and was most recently a fellow at NYC Data Science Academy. Patrick is always keeping up with the latest machine learning techniques in astronomical and public policy research.