September meetup

Join us for another AI-focused meetup at AI Sweden, Lindholmsplatsen 1! We’re bringing together the community for an evening dedicated to production-grade AI infrastructure, exploring how to run local AI models on Kubernetes, manage GPU resources effectively at scale, and secure your data sovereignty without getting trapped in closed SaaS ecosystems. Special thanks to AI Sweden, Conoa, and Proact for sponsoring the venue, food, and drinks for this event!

Agenda

• 17:30 – 18:00 | Doors Open, Food & Networking - Grab a bite to eat and something to drink. Settle in, connect with peers, and get ready for the evening's talks.
• 18:00 – 18:05 | Welcome & Introduction - Brief opening remarks by the Cloud Native Göteborg organisers team and our host, AI Sweden.
• 18:05 – 18:35 | Sovereign AI on Kubernetes: Why and How to Own Your AI Stack Speaker: Kenneth Albinsson, CTO, Conoa Abstract: Relying entirely on closed cloud services and proprietary models like Claude and GPT-4 introduces strategic risks around data sovereignty, regulatory compliance, and vendor lock-in. Kenneth shares a strategic perspective on why organizations are moving toward running local open-source models on Kubernetes, outlining the architectural vision, cost trade-offs, and governance frameworks needed to maintain complete control over your AI future.
• 18:35 – 18:40 | Short Break
• 18:40 – 19:10 | Running GPU on Kubernetes Speaker: Andreas Lindhé, Platform Engineer, Conoa Abstract: Andreas will show how they use GPUs in Kubernetes at SciLifeLab and KTH. It's a practical demonstration of how the NVIDIA GPU Operator can be used to enable Pods to use GPU resources, some examples of how to use it and discussions about issues along the way.
• 19:10 – 19:30 | Sovereign AI Ops: The Self-Healing LLM Cluster Speaker: Karl Malmros, Platform Engineer, Conoa Abstract: Demonstrating a live scenario where an Open Source LLM agent autonomously monitors its own GPU infrastructure. How does the AI detects failures locally, without sending data to the cloud, and heals itself in real-time? See how to build a closed-loop system where your AI stack is not just compliant, but truly self-sufficient.
• 19:30 – 20:30 | Networking & Post-Talk Mingling - Continue the conversation with the speakers and fellow attendees.