Tech Meetup: GPU-Management, Kubernetes, Multi-Cloud
Details
On the agenda
🗣️ Topology-aware cloud compute: Building predictable GPU instances on Kubernetes
Speakers:
Lukas Stockner & Peter Bangert, Platform Engineers at Stelia.
Abstract:
Modern GPU workloads do not just need “some CPUs” and “some memory”; they need resources that are physically close, isolated, and predictable. In this talk we show how Stelia builds topology-aware cloud compute on Kubernetes, mapping Linux CPU, memory, NUMA, cgroup, and device information through Kubernetes scheduling and into QEMU so GPU instances behave like the hardware they were promised.
🗣️ MetaCloud: beyond multi-cloud. Turning fragmented infrastructure into one programmable resource
Speaker: Gabriele Fronzé, CEO and Co-Founder of Elemento Cloud, democratizing cloud computing with vendor-neutral solutions that provide greater control, lower costs, and reduced environmental impact.
Abstract: Multi-cloud promised freedom and delivered fragmentation: every provider with its own APIs, its own economics, its own lock-in. Its own UI. This talk introduces the MetaCloud, a vendor-neutral layer that turns heterogeneous public providers, private hardware and edge sites into a single programmable pool. A pool of resources handled via an output-first intent layer. Covering what it is, why it is becoming necessary, how it works in practice, and what it changes in terms of control, cost and environmental impact.
Moderation: Melody Sofia Eroshevich, Site Reliability Engineer & Community Lead at CLOUDETEER
Who should attend?
This event is designed for CTOs, CIOs, DevOps professionals, platform engineers, and anyone working with modern infrastructure.
Get ready for two exciting tech talks, complete with live demos and plenty of time to connect, exchange ideas, and dive deeper with industry experts. Snacks & drinks included.
No sales pitch. Just tech ‼️
Discover in the video what tech meetups with Leaseweb feel like