KubeFlow: Pythonic Machine Learning at Scale on Kubernetes | SciPy 2018 | Bailey & Aronchick

“KubeFlow marks the beginning of the end of the data scientist and/or software engineer as disparate roles. Like DevOps has merged operations and development, DataDevOps will consume data science.” - Philip Winder, an engineer and consultant at Container Solutions

http://container-solutions.com/tensor...

You've created and tuned a machine learning model, using TensorFlow, PyTorch, or scikit-learn - now what? How can you ensure that the model is deployed to your DevOps team as production-ready code, and can scale as needed on incoming data? How can you seamlessly migrate a model from your local laptop / virtual machine to a hosted server on your cloud platform of choice?

This talk walks through the architecture of Kubeflow: a project dedicated to answering those questions - and to making machine learning on Kubernetes simple, portable and scalable.

We will describe, in detail, the three components of the project:

a JupyterHub platform for creating and managing Jupyter notebook servers that are used by data science and research groups;

a Tensorflow Customer Resource for managing compute resources to a specific cluster size; and

a Tensorflow Serving container to house the machine learning work.

By the end of this talk, you will have a firm understanding of why Kubernetes would be useful to machine learning engineers, and how you can deploy it, today, to support your predictive models.

https://github.com/kubeflow/kubeflow

See the full SciPy 2018 playlist at    • SciPy 2018: Scientific Computing with Pyth...