Toronto Apache Spark #10

Title: Scalable Item Based Collaborative Recommendations with PySpark

Agenda:

6:30PM to 7:00PM - Opening and networking (refreshments provided)

7:00PM to 8:30PM – Re-engagement Targeted Item Recommendations

8:30PM to 9:00PM - Networking

Broadcast Link: https://hangouts.google.com/call/t7k2wzz2wndn5hrqipn2ps46cye (https://plus.google.com/events/c7tmvquc552fifb0j5s34h3vs34)

Speaker(s):

Mo Kobrosli (https://www.linkedin.com/in/mohammed-kobrosli-b9a42643) from eBay Classifieds Finding/Search Science Team (FiSci)

The FiSci team is responsible for the research and development of solutions to tough problems in the finding domain that cut across different platforms and countries within the eBay Classifieds portfolio (http://www.ebayclassifiedsgroup.com/) (e.g. Kijiji Canada, Gumtree UK, Markplaast NL, etc…)

Description:

One of our recent problems? Coming up with a scalable and effective model of producing re-engagement targeted item recommendations to buyers based on their recent interactions on site. The challenge? Short lived volatile content with the major of buyer sessions occurring anonymously.

Level: Intermediate/Advanced

Target Audience: Data Scientist, Data Engineer

Sponsor:

http://photos3.meetupstatic.com/photos/event/8/a/a/600_449882218.jpeg

Organized by Marshall Berenbaum, Mehrdad Pazooki