Ask not what you can do for your platform - ask what your platform can do for you!

Welcome

Platform engineering is no longer just about standardizing templates. It’s about closing the gap between raw infrastructure and developer velocity, no matter where that infrastructure lives. And making the outcome digitally sovereign in one washup.

Tonight, we’re charting that entire spectrum. We’ll explore how to coordinate complex cloud-native workflows effortlessly with OpenChoreo, bypass deployment friction entirely with Codesphere’s instant, sovereign development environments, and push the absolute limits of edge portability by running Kubernetes directly on a mobile phone. From seamless orchestration to clusters that fit in your pocket. Tonight is all about redefining what a platform can do for you.

"My fellow CNCFitzens: ask not what you can do for your platform - ask what your platform can do for you." Overheard at a last KubeCon, Anonymous

📍 Coordinates: BioQuant, Grosser Hörsaal (Raum 041)

🗓 Agenda:

• 🕒 18:30 – Check-in
• 🕒 18:45 – Welcome - Cloud Native Organizers Heidelberg
• Talks
• 🎨 21:00 – Networking. Use the opportunity to get to know somebody new!

Talk 1 (Virtual): Five Years of Rebuilding the Same Kubernetes Platform

• Abstract: Every platform team I've met has built the same thing. A portal, a CI pipeline, some GitOps wiring, an observability stack, and RBAC glue holding it together. Then they own it forever, and usually three people understand it. We did this at WSO2. We built Choreo, ran it as a SaaS for years, and hit every wall you'd expect. Then we rebuilt it in the open as OpenChoreo, which joined the CNCF Sandbox and shipped 1.0 this March. In this talk I'll go through what we got wrong the first time and what we changed. Why the control plane never gets a kubeconfig for your workload clusters. Why we froze releases into immutable snapshots instead of re-rendering from live state. And why platform engineers configure golden paths in validated schemas rather than writing their own controllers. If you've built this glue yourself, come argue with me afterwards.

• Speaker: Isala Piyarisi. Isala is an Associate Technical Lead at WSO2 and a maintainer of OpenChoreo, the CNCF Sandbox open-source internal developer platform. He got into infrastructure at 12, defending a DDoS'd Call of Duty 4 server, and later built WSO2's eBPF observability layer before turning to OpenChoreo.

Talk 2 (in person): Beyond Hyperscalers: Architecting a Production-Grade Internal Developer Platform on CNCF Primitives

• Abstract: Modern developer platforms often come with a hidden catch: hyperscaler lock-in or massive operational overhead. To give our engineering teams true self-service speed without compromising infrastructure freedom, we built an Internal Developer Platform (IDP) that runs anywhere—from on-premise bare-metal to any cloud provider. In this talk, we’ll dive under the hood of our platform built entirely on Kubernetes and open-source CNCF projects. We’ll show how we orchestrate entire service landscapes, integrate web-based IDEs, terminal access, and monitoring cockpits, and provide an extensible internal service marketplace. We’ll also talk about handling heavy infrastructure requirements—up to bringing entire storage clusters like Ceph along when the environment demands it. Most importantly, we’ll share the unfiltered lessons from running this setup in production within the German tech stack (DE-Stack), how we navigate vendor-neutral platform engineering, and how we leverage and contribute back to the open-source ecosystem.
• Speakers: Alex Klein and Utku Erol. Alex is a tech lead at Codesphere, leading the team responsible for the platform's core deployment and monitoring capabilities. Before, he was part of a team building an internal cloud platform on on-prem bare-metal servers. In his free time he enjoys being outdoors, hiking, and climbing. Utku joined Codesphere as a tech lead, to build out its extensible service marketplace. He utilizes his experience of having built multiple microservice architectures. At SAP, he was part of the observability team. He has a beagle called Spike.

Talk 3 (in person): Clusters that fit in your pocket

• Abstract: Running Kubernetes on a Phone. As smartphones are becoming more and more powerful, do you ever wonder whether you can use smartphones as Kubernetes nodes and use the AI-capabilities of the devices in your application? In this talk I'll show you how this works with Android 16 and a few bash scripts

• Speaker: Johannes Bechberger. Johannes is a JVM developer working on profilers and their underlying technology in the SapMachine team at SAP. This includes improvements to async-profiler and its ecosystem, a website to view the different JFR event types, and improvements to the FirefoxProfiler, making it usable in the Java world. His work today comprises many open-source contributions and his blog, where he regularly writes articles on profiling and eBPF.