Kubeflow: Machine Learning Over K8s

• The meetup will be online and in Hebrew
• Link to the event will only be visible to people who fill out the form - https://bit.ly/kubeflow-meetup-reg
• The online meetup will be held on July 7th, 7pm, IL Time Zone (GMT +3:00)

Agenda:
19:00 - 19:15 - MLOps Challenges
19:20 - 19:50 - Intro to Kubeflow
19:55 - 20:30 - Simplifying and extending Kubeflow with serverless ML functions
20:30 - 21:00 - Q&A with panelists

// Intro to Kubeflow
As the involvement of Machine Learning components in software products is increasing rapidly, there is a growing need for tools to operate and orchestrate Machine Learning development and deployment. One of those tools is Kubeflow. Developed by Google for the Open Source community, this tool aims to organize the cycle of model development, testing, and deployment on the infrastructure of Kubernetes. In this talk, we will introduce the tool, its capabilities, and its integration in the infrastructure of organizations.

Read about Haim's experience with Kubeflow in his post - https://medium.com/everything-full-stack/is-kubeflow-designed-for-data-scientists-yes-but-1110c8fb35ec

Haim Cohen // Bio
Haim Cohen works as a Machine Learning and BigData engineer at Tikal, helping companies implementing infrastructure for large Machine Learning and BigData projects. As such, Haim reviews and tests various tools to help Data Scientists develop and deploy production-level systems incorporating ML models.

The process of moving from data-science research to production pipelines is long and resource-consuming, new practices like MLOps and tools like Kubeflow (ML toolkit and pipeline management over Kubernetes) are emerging to provide the equivalent of CI/CD for data science projects, but this requires dedicated ML engineering teams to translate data-scientists/engineers work to production-ready code.

// Simplifying and extending Kubeflow with serverless ML functions

Serverless can simplify data science by automating the process of code to container and enables users to add instrumentation and auto-scaling with minimum overhead. However, serverless has many limitations involving performance, lack of concurrency, lack of GPU support, limited application patterns and limited debugging possibilities. Yaron Haviv will introduce Kubeflow, and how it works with Nuclio and MLRun, open source projects enabling serverless data-science and full ML lifecycle automation over Kubeflow.

Yaron will show real-world examples and a demo and explain how it can significantly accelerate projects time to market and save resources.

Yaron Haviv // Bio
Yaron Haviv is a serial entrepreneur who has deep technological experience in the fields of data, cloud, AI and networking. As the CTO of Iguazio, Yaron defines the company’s vision and strategy for the company’s data science platform and headed the shift towards real-time AI. He also initiated and built Nuclio, a leading open-source serverless platform with over 3,000 Github stars. Prior to Iguazio, Yaron was the Vice President of Datacenter Solutions at Mellanox, where he led technology innovation, software development and solution integrations. He was also the CTO and Vice President of R&D at Voltaire, a high-performance computing, IO and networking company. Yaron is an active contributor to the CNCF working group and was one of the foundation’s first members. He presents at major industry events and writes tech content for leading publications like TheNewStack, Hackernoon, DZone, Towards Data Science and more.

#backend