London Platform User Group (LOPUG), Wednesday 8th July, 6.30pm onwards
Details
For our July meetup, we have a change of venue and will be hosted by the lovely folks at Adaptavist. We have two excellent talks lined up and the agenda will be:
- 6:30pm - arrive, drinks, networking
- 7:00pm - Welcome & Housekeeping
- 7:15pm - Talk 1: "Stop Wasting GPUs: How We Built a Golden Path for GPU Sharing on Kubernetes"
Everyone is hyped about frontier models as a service. The moment you need to run them yourself, for latency, cost, compliance, or just control, the bill shows up. You pay for a whole GPU even when your pod uses a sliver of it. Kubernetes has always counted GPUs like bricks: one pod, one card, done.
That just changed. Kubernetes 1.34 ships a new way to schedule hardware, and NVIDIA lets you physically slice one card into isolated partitions. Put them together and you get real GPU sharing on Kubernetes. No time-slicing tricks. Real isolation, real savings.
We built the whole stack from an empty cluster. Come to the talk and watch it run end to end on stage.
Speaker: Engin Diri, Senior Solutions Architect @ Pulumi
As a Senior Solutions Architect at Pulumi with over 15 years of experience in the IT industry, including roles at the Schwarz Group and W&W Versicherungen, I bring extensive expertise with an end-user and enterprise focus. Currently working for a startup while collaborating with enterprise clients has further enriched my experience!
- 7.45pm - food, more drinks & networking
- 8.00pm - Talk 2: "Bringing Back the Feedback Loop: Platform Engineering for AI-Assisted Delivery"
What changes when AI starts participating in delivery systems, and how can platform teams keep feedback loops, governance, and developer experience intact when AI is making everything happen faster than ever before? Matt will make the case for not throwing away our old practices in this new era, and for bringing back DevOps-style feedback loops stronger than before.
Speaker: Matt Saunders, VP DevOps at Adaptavist
Matt Saunders is VP DevOps at Adaptavist, leads the DevOps section of InfoQ, and co-organises the London DevOps AI meetup
- 8.30pm onwards - drinks and networking.
So, please come and join us, we look forward to seeing you there.
The LOPUG team.
- Food and drinks provided
- Good time will be had by all