Machine Learning in the cloud with Mahout and Whirr

These days Apache Mahout becomes an increasingly popular framework to do large scale Machine Learning on massive amounts of data. In order to get the maximum out of Mahout, it requires to run on large computer clusters running Apache Hadoop. Unfortunately deploying and maintaining such clusters and the various Mahout jobs comes with a large overhead which diverts much of the focus to administrative tasks. The Mahout Whirr service was created to put this focus back to the your data and the different algorithms running on it. In this talk I give an introduction to Apache Mahout and Apache Whirr. Additionally I show you the Whirr’s Mahout service, how it is implemented and discuss some tips and tricks when working with these technologies.

Watch the video of Frank Scholten talk here.

Slides: 

Machine Learning in the Cloud-fscholten-bbuzz12.pdf