Deploying Containerized AI/ML Workloads Across Edge-Core-Cloud with openSUSE
Rudraksh Karpe
Rudraksh Karpe, a Google Summer of Code 2024 contributor at openSUSE, is passionate about AI/ML and the Cloud-Native space. He is determined to contribute to open-source software and give back to the community. During his early years, switching between operating systems ignited a spark of curiosity within him to explore technology that has continued to burn bright. He has been contributing to the Analytics Edge Ecosystem Project by openSUSE and Rancher, building ML solutions on the Edge for various business verticals.
No video of the event yet, sorry!
Deploying containerized AI/ML workloads across edge, core, and cloud infrastructures is crucial for optimizing data processing, enhancing performance, and managing complex computations efficiently. This talk highlights our practical experience from the Analytics Edge Ecosystem project during Google Summer of Code 2024, demonstrating how openSUSE Leap can be leveraged to achieve these goals. By utilizing Kubernetes and lightweight distributions like K3s on edge devices, we have reduced latency and minimized bandwidth usage, while hybrid and multi-cloud deployments offer scalability and flexibility.
In this session, we will explore the technical aspects of deploying various AI/ML workloads at the edge using openSUSE Leap.
Key topics:
- Using openSUSE Leap as base layer for deployment using KVM
- Containerization of workloads with Podman/Docker
- Kubernetes orchestration and cluster management with Rancher by SUSE
- Implementing a data pipeline that seamlessly transfers data from Edge devices
- Using AI/ML models deployed at the edge to provide real-time insights
- Cost optimization using edge devices for local processing
- Ensuring data security by processing sensitive information locally at the Edge
- Date:
- 2024 November 2 - 16:30
- Duration:
- 40 min
- Room:
- Room A
- Conference:
- openSUSE.Asia Summit 2024
- Language:
- en
- Track:
- openSUSE
- Difficulty:
- Medium
- Who broke the build? — Using Kuttl to improve E2E testing and release faster
- Start Time:
- 2024 November 2 16:30
- Room:
- Room B