Deploying Machine Learning Models as Microservices Using Docker

Modern cloud applications often rely on REST-based microservices architectures using Docker containers. Docker allows your applications to communicate with each other and to create and scale different components. These techniques are used by data scientists to efficiently scale machine learning models to production. This video teaches you how to deploy machine learning models behind a REST API--to serve low latency requests from applications--without using a Spark cluster. You'll also learn how to export SparkML models; how Docker can be used to build, deploy and ship microservices code; and how a model scoring system should support both single-on-demand predictions as well as bulk predictions. Basic knowledge of Scala, Python, Hadoop, Spark, Pandas, SBT or Maven, cloud platforms like Amazon Web Services, Bash, Docker and REST is required.

Mikhail Semeniuk works as a data engineer at Shift Technologies. Mikhail was a senior-level statistician at UnitedHealth Group for six years. It is the largest U.S. health insurance provider. He has a BS degree in Economics and Financial Math from University of Minnesota.

Jason Slepicka, a senior data engineer at DataScience, is Jason. Jason is currently pursuing a PhD in Computer Science at University of Southern California Information Sciences Institute.