Agentic workloads require sovereign control planes to ensure compliance and avoid vendor lock-in. This session details building a vendor-neutral AI control plane on Kubernetes. We examine a stack covering orchestration, API gateways, observability, QA, and security-as-code. Using GitOps, we’ll demonstrate how to manage agent lifecycles, ensuring the infrastructure is as autonomous as the agents it hosts. Learn to architect a scalable, portable AI platform that prioritizes digital sovereignty without compromising technical excellence.

Target Audience: Platform Engineers, Software Architects, AIOps
Prerequisites: Basic knowledge on cloud infrastructure
Level: Advanced

Extended Abstract:
Agentic workloads demand more than just GPUs; they require a sophisticated, sovereign control plane. To remain competitive and compliant - particularly in the EU - organizations must move away from proprietary silos toward vendor-neutral, open-source architectures.

This session breaks down how to build a sovereign AI control plane from the ground up using Kubernetes. We will examine the stack required for agentic success: orchestration, API gateways, observability, continuous quality assurance, security and compliance-as-code. We’ll demonstrate how GitOps principles can be applied to manage agent lifecycles, ensuring your infrastructure is as autonomous as the agents it hosts. You will walk away with a clear understanding of how to architect a scalable, portable AI platform that prioritizes digital sovereignty without compromising on technical excellence.