Building a Scalable Data Science & Machine Learning cloud using Kubernetes BigData и машинное обучение

Доклад отклонён
Murali Paluru
Rancher Labs

Murali Paluru, International Conference Speaker, Principal Software Engineer at Rancher Labs, Inc, started his career in networking in the mid-2000s and after working on various products lines at different companies, he is currently developing solutions involving containers and Kubernetes. Some of his past projects involve enabling secure connectivity between containers as part of multi-cluster federation, implementing network policies, CNI plugins, enhancing network performance and stability. While not hacking stuff using the Go programming language he loves to interact with other developers at meetups, conferences to exchange ideas, share his experiences and lessons learnt along the way.

https://www.linkedin.com/in/leodotcloud
https://twitter.com/leodotcloud
https://github.com/leodotcloud
Тезисы

A real-life story of architecting & building a cloud-native data science platform using Kubernetes.

A growing team of data scientists was looking for a simple, flexible, scalable and a secure way of migrating to the cloud as the on-prem data center started becoming a bottleneck. Kubernetes, which enables applications to run on a variety of private and public clouds, along with an ever-growing feature set, matched most of the team's requirements.

In this talk, Murali Paluru, who had the opportunity to work with the Data Scientists team, will share the gathered requirements, architecture and in-depth details of the Data Science platform built on top of Kubernetes. He will also demo a one-click solution that he has developed, to hide away the complexity of Kubernetes and enable the Data Scientists to focus on analyzing the data instead of managing the underlying infrastructure.

Data Science is gaining much momentum and is being used by various organizations to get insights into the troves of data being accumulated. The attendees will learn how they can leverage Kubernetes within their teams. They can use the architecture shared in this presentation as-is or build on top of it and they can avoid the common mistakes and pitfalls that would be shared in this presentation.

Архитектуры / другое
,
Devops / другое
,
Инфраструктура как сервис (IaaS), платформы как сервис (PaaS)
,
Big Data и Highload в Enterprise

Другие доклады секции BigData и машинное обучение