Workshop: DevOps in Machine Learning: from Development to Production in 90 Days


If you had 90 days, what tools could you leverage to develop, train and deploy a machine learning model into production? In this talk I would like to share the story of how we developed JibJib, a machine learning powered app to classify birds by their calls, our entry for the Coding Davinci Ost 2018 Hackathon. Using tools like Docker, SaltStack and Terraform we have built a microservice architecture that supports a model hosted via TensorFlow serving, Google’s open-sourced solution for running machine learning in production. I hope to give an overview of what kind of problems have to be dealt with when creating a machine learning powered service intended for production use, as well as how DevOps toolchains can help solving them.


Alexander J. Knipping

I’m currently a student at TU Chemnitz, writing my Masters thesis on the orchestration of microservices. I’m absolutely passionate about DevOps and have been doing little than delving into the depths ...