
Speaker "Diego Klabjan" Details Back

-
Name
Diego Klabjan
-
Company
Opex Analytics
-
Designation
Founder
Topic
Personalized recommendations from web events in Spark: Challenges and opportunities
Abstract
On a web portal users bid for projects, each one – users and projects – having many features. Due to a large number of projects at any point in time, only specific projects should be recommended to a user to potentially bid on on a daily basis. Due to the large size of underlying data and many models to solve, Spark with its powerful machine learning library has been selected. Our experience and enhancements made to Spark will be discussed. The audience will learn about challenges and expertise required in Spark for a problem that is not from a textbook.