If you're interested in microservices, text search, or data consistency strategies in the wild then come join us for 3 short talks on Tuesday, June 20th, at the Quizlet office! These talks are intended to give you a chance to understand the design and implementation of solutions employed by companies with high throughput data platform requirements. We hope you'll learn new things, ask questions, and meet other people at the event.
The event starts at 6:00 pm at the Quizlet office, located at 501 2nd St. in Soma (At the corner of 2nd and Bryant). We'll have food and drinks before starting the talks at 6:30 pm and additional time for questions and discussion afterwards. Each talk will be 15 minutes with time for questions afterwards.
Introducing Tripod: Flickr’s Backend, Refactored Peter Welch (Senior Principal Engineer, Flickr)
Peter recently led the re-architecture of Flickr’s back-end services to create a new platform called Tripod, which is used at large scale for Yahoo Mail and other consumer services. This talk will cover building Tripod APIs in Yahoo's global context required a deep focus on eventual consistency challenges, sparse APIs and content deduplication. The team also had to figure out how to build an active/active system to ensure high availability.
Peter Welch is a Senior Principal Engineer at Yahoo, where he uses his 20+ years of experience to lead their media services platform. Prior to Yahoo, Peter led the Flickr engineering team in building Flickr’s Magic View which uses computer vision to classify and organize 30 billion photos in real time. Prior to Yahoo, Peter served in engineering leadership roles at Deem, Blue Martini, Kana, and other companies in the Bay Area.
Searching for the Best Learning Content Karen Sun (Engineering Manager, Quizlet)
Quizlet is the most popular online learning service in the US and has amassed over 5 billion user generated flashcards in dozens of languages across all areas of study— medicine, math, foreign language study, standardized tests, and more! Quizlet leverages this enormous bank of knowledge to connect users with highly corroborated and customized learning material for any concept they need to master. We will discuss the strategies we've employed to tackle both global subject-level document search and aggregation based concept-level search on top of Elasticsearch. We'll cover lessons learned related to optimal document structure, pre-processing strategies, sharding configurations, and cluster stability.
Karen Sun leads the Backend/API and Data Science teams at Quizlet, focusing on building their data platform to support studying, teaching, and the cognitive science research that powers all the app experiences.
Tenements in the Troposphere Scott Maher (Sr. Tech Lead, Twilio)
Distributed systems built on top of the cloud need to consider the material of their foundations; ever changing and susceptible to chaotic environmental effects. This talk is about building microservices for reliability when downstream components will be flaky or unpredictable. We will discuss managing fairness and reliability with techniques like quotas and bulkheading in multitenant async IO heavy services and how we build them with Akka Actors, Akka Streams, and Scala.
Scott Maher is a senior tech lead at Twilio. He previously led development for their SMS and MMS platforms, and is now connecting Twilio with dozens of OTT messaging and IoT networks like Facebook Messenger and Amazon Alexa.