[Pycon] [new paper] "Akash Tandon" - Kubeflow: portable and scalable machine learning on top of Kubernetes

[info a pycon.it]

Title: Kubeflow: portable and scalable machine learning on top of Kubernetes
Duration: 60 (includes Q&A)
Q&A Session: 15
Language: en
Type: Talk

Abstract: ML solutions in production start from data ingestion and extend upto deployment. We want this workflow to be scalable, portable and simple. Containers and Kubernetes are great at the former two but not the latter if you aren't a DevOps practitioner. It is safe to say that most of the data science community isn't. Kubeflow is an open source project that aims to bridge this divide.

Kubeflow is a cloud-native ML platform built on top of Kubenetes and based on Google's internal ML piplelines, In this talk, we'll explore how you can leverage it to deploy best-of-breed open-source ML systems to diverse infrastructures. 

Preliminary outline
- Need for composable, scalable and portable data science and ML
 - Kubernetes for ML: challenges and opportunities
- Jupyterhub and notebooks: an overview
 -Tensorflow and other supported ML frameworks: an overview
 - Kubeflow: ML toolkit for Kubernetes
 - Who is it for?
 - Live demo

Tags: [u'Reproducibility', u'kubern', u'devops', u'tensorflow', u'machine-learning', u'DataPipelines', u'open-source', u'CloudComputing']