Apache Pinot Virtual Meetup with LinkedIn

Link: https://linkedin.zoom.us/j/96049868422?pwd=bFFBeGlSRWFNUnpUeEYxdnE1bDluQT09

Meeting ID: 960 4986 8422
Password: 903807

We welcome you to join us for a virtual meetup hosted by LinkedIn and the Apache Pinot community. After the success of our last meetup in May with Uber, we've seen an influx of new users and community members looking to explore Apache Pinot.

Agenda:

6:00 PM — Welcome
Join the virtual meetup link and introduction.

6:10 PM – Multi-Set Count Distinct Analytics using ThetaSketches in Pinot
As LinkedIn continues to grow, there is an increasing need for approximating large multi-set distinct-count analytics queries. This talk will cover how LinkedIn is using data sketching algorithms to approximate large set cardinalities with complex boolean expression queries.

6:30 PM – Scaling Pinot for LinkedIn's Feed
LinkedIn Feed, the landing page for members, is one of the most popular features, and therefore maintaining feed quality is vital for member experience. Every member action on feed posts/articles (view/like/share/comment) are ingested into Pinot in real-time. This data is in turn used in ranking feed articles to show to the user next time. Come join us in this talk to learn how we scaled Pinot to serve 11k qps at p99 of 50ms, while ingesting 50k events per second, to build such a tight real-time feedback system at scale for 700M members.

Speaker — Seunghyun is a Senior Software Engineer at LinkedIn's Big Data Platform team. He is an active committer to Apache Pinot (incubating). He has worked on Pinot's core features including replica group aware routing and segment merge and roll-up. His main interest is solving all kinds of problems in distributed systems.

6:50 PM — Application & Tuning Apache Pinot for Personalization usecase
Come on a journey with Publicis Sapient's team where we explore the various facets of Apache Pinot and how we tuned the OLAP datastore to make sure it performs best to ensure the synergy of the complete eco-system. Learn how we applied different mechanisms and reduced the response time of Apache Pinot from 100s of milliseconds to 10s of milliseconds and that too with a 10-fold increase in throughput.

Speaker — Srisudha Garimella works as a Manager, Technology in Publicis Sapient. Sudha has proven expertise in using Adobe Marketing Cloud Tech stack, designing and building distributed systems using Microservices, Apache Kafka and other open source tools. She enjoys working in new challenges because it brings her back to what she enjoys best -- learning along the way & writing code. Additionally she leads different teams to solve business problems.

7:10 PM — Closing and Q&A
Closing remarks and audience Q&A