Export or edit this event...

CNPDX June: Machine Learning and Example Kube Apps

221 Southeast Ankeny Street
Portland, OR 97214, US (map)



Autodesk is hosting us for the first time for an application-centric meetup! Learn how to build Kubernetes projects. First we'll show you an end-to-end example Kubernetes application using Helm, and then we'll show you a Machine Learning/NLP project.

PLEASE RSVP on Meetup.com if you have an account for food ordering reasons.


Devstats: An example end-to-end Kubernetes application.

Presenter: Josh Berkus, Red Hat

You may know DevStats as a repository of contributor statistics for Kubernetes (devstats.cncf.io), but you're about to find out that it is also a great demonstration case of migrating a complete application to orchestrated microservices. Our community has lacked solid, production-quality, end-to-end, all-open-source application examples, so we decided to make DevStats into one.


NLP for fun and profit. Story of end to end machine learning project on Kubernetes.

Presenters: Michał Jastrzębski, Hamel Husain, Github

Abstract: Machine learning makes waves again, we are being constantly amazed by new feats neural networks can achieve. Models, while important, are a small part of the whole machine learning system. The infrastructure it runs on, the dataset itself and data cleaning/preprocessing pipeline are often omitted in news articles, but they are critical components of any machine learning project. Infrastructure, in particular, is a hard problem and there are few good examples of running cloud-native machine learning. In GitHub, we are building a full platform for machine learning based on Kubernetes. In this talk we will walk through such end to end project, one anyone can replicate at home (code is open source and data is available), we will tackle hard natural language problem - automatically labeling GitHub issues as bug or feature. We will deep dive into our Kubernetes use case and walk through different components we have used for every step of ML project lifecycle - from data preprocessing to serving model for the application.