Lifting AntiSpam to the cloud and beyond

Stream
06/11/2018 - 11:50 to 12:10
Maschinenhaus
short talk (20 min)
Intermediate

Session abstract: 

 

Throughout the last months, our AntiSpam platform underwent many significant changes. The biggest of them, the migration to the cloud, and the removal of our Hadoop cluster made us change the way we manage and work with our data.

This talk will be about this transition. During the first part of the presentation, I will focus on the reasons why we performed the migration and the architecture of our system. Particularly, I want to share which components of Google Cloud we are using, and most importantly how we use them.

The second part will be dedicated to the small improvements and advantages we got from the cloud. For example, how BigQuery made it easier for us to train our machine learning models, and how we use Data Studio to make sure our predictions are on point.

Lastly, I will conclude by introducing the new techniques and machine learning models we have developed to detect and punish spammers since our last time at Buzzwords '17.

 

Video: 

Slide: