Running ML Infrastructure on HBase

Fred SadaghianiSeptember 23, 2014

We recently¬†hosted our first ever HBase meetup! This was a very exciting event for us as it was the first time we showed off some of the great infrastructure and systems we’ve built to power our machine learning platform.

Of course, we didn’t start with HBase. When we first launched in April 2012 our platform was built on MongoDB. At the time, Mongo provided a great balance between flexibility and operability, but we very quickly outgrew it and moved to HBase and now proudly serve thousands of sites and many 10s of thousands of requests per second on our HBase cluster.

In our talk at the Meetup, Andrey focuses on the underlying infrastructure we have built to support both online and offline learning at scale and how HBase, in particular, lends itself to this problem.

We look forward to hosting more meetups around infrastructure, systems and data science in the coming weeks and months. If you’re interested in learning more or hacking on HBase and machine learning, please don’t hesitate to reach out to us!