Dead Simple AB Testing at Massive Scale with Scala and Spark

Sean Quigley | GIPHY

ABOUT THE TALK

As a small engineering team, GIPHY needs creative, lightweight solutions to deliver features for a massive global user base. Leveraging the type-safety and expressiveness of Scala along with the performance of Spark allowed us to rapidly deliver a reliable AB testing solution for our search engine. We'll discuss GIPHY's cloud-based data architecture, our Bayesian AB testing framework, and insights from using it in production for the past 8 months. In particular, we want to focus on various engineering principles, such as compile-time guarantees and a monorepo codebase, that we believe enable small engineering teams to continue to focus on delivering product features as a business scales.

Download Slides

sean quigley

Platform Engineer | GIPHY

Sean has always been fascinated by the abstract process of solving logic-based problems. This is the common theme on his journey from a fondness for math and puzzles as a child, to studying Physics and Economics at the University of Michigan, working as a hedge-fund quant, a data scientist in ad-tech, and now a software engineer developing solutions for web-scale analytics and search at GIPHY. In his spare time, he enjoys martial arts, electronic music, and traveling.

Sean Quigley