We will share with you how we built an Automated Machine Learning platform upon Openstack designed to handle scalability and automation.

Our first decision was to use Openstack for the compute and Swift for the storage. On the compute side, Nova allows us to deploy different flavors in order to handle specific use cases, like GPU for TensorFlow training. On the storage side, Swift was the perfect solution to abstract the storage of each artifact created by the platform.

During this talk, we will discuss about some use cases solved by the Machine Learning platform such as predictive maintenance on our infrastructures.

We will outline the architecture of the project, built on Open Source software including Spark, CDS, Warp 10, PMML, Scikit-Learn and TensorFlow. We will also explain how OVH is laying the next steps of a fully open Source Automated Machine Learning platform allowing Data Scientists to work on Stage-Gate innovation processes and go to production.

Stream Summit keynote videos including case studies from China Mobile, China Telecom, China Unicom, Baidu, SK Telecom and CERN

Search Result

Schedule Matches

Speaker Matches