Platform Engineering Linz

We will have the first Platform Engineering Linz Meetup in January 2026. We are still preparing the talks and we look forward seeing you there. Until then, have nice holidays and see you in Linz :)

In this event, you will learn about the dual observability needs of Internal Developer Platforms (IDPs) and how improving platform observability enhances application performance, practical strategies for mentoring junior engineers by bridging the gap between idealized learning paths and real-world challenges, and the process of building a secure, in-house LLMaaS platform for government use, focusing on architecture, security, and implementation

Agenda:

Doors open: 5:30
-----------------------------------------------------------------------
Official start: 6:00 - sponsor talk 5 mins
Talk 1: --> 6:30
Speaker: Andi Grabner
Topic: Stranger Platform Engineering: The Two Sides of Observability in Your IDP

Abstract: Your Internal Developer Platform lives in two worlds. In the Right Side Up, developers rely on self‑service observability to build, deploy, operate and debug their apps. In the Upside Down, the platform itself must be observable to ensure its reliability, performance, and to understand usage patterns and adoption.
The twist? these two worlds are not separate. Improving observability of the platform directly improves the observability for the applications built on top of it. Just like in Stranger Things, observability becomes the force that connects both realities—creating feedback loops, strengthening the platform’s foundations, and enabling teams to spot issues before they cross between worlds.
This talk reveals why mastering both sides of observability is essential for any modern IDP. Expect real-world examples, a few supernatural analogies, and practical guidance you can bring back to your platform team—no Demogorgon hunting required
-----------------------------------------------------------------------
Break to switch speakers: --> 6:40
-----------------------------------------------------------------------
Talk 2: --> 7:00
Speaker: Katharina Sick
Topic: Turning Platform Engineering Right Side Up: Better Learning Paths for Juniors

Abstract: Breaking into platform engineering and DevOps often feels like navigating two extremes. When I started, I bounced between perfect, predictable, and easy to follow sunshine tutorials, and the messy realities of production, where overwhelming systems and uncertainty left me stuck and, at times, spiraling into procrastination. That gap shaped my early years and still shows up when I mentor today.It’s a bit like stepping into the Upside Down: confusing, chaotic, and full of unexpected challenges. The real issue isn’t just complexity. Tools are often taught in isolation, and juniors are left figuring out how they connect in practice. With countless combinations, there’s no one-size-fits-all guide.This talk explores what I have found useful when mentoring newcomers to the platform engineering space. We’ll look at ways to make connections visible, create safe spaces to apply knowledge, and build community support, so engineers can grow without being overwhelmed and teams can onboard more effectively.Whether you’re a junior yourself or someone looking to support the next generation, this talk offers actionable ways to turn frustration into meaningful progress.
-----------------------------------------------------------------------
Break: --> 7:30
-----------------------------------------------------------------------
Talk 3: --> 8:00
Speaker: Christian Kirchknopf
Topic: Building a Secure In-House LLMaaS Platform: BRZ's Approach to AI for Federal Government

Abstract: In this talk, I demonstrate how we (BRZ) have developed an Enterprise-Ready LLM-as-a-Service platform that provides government applications simple and secure access to Large Language Models – without dependency on public cloud providers and without data sharing with third parties.
Through a unified API, AI applications gain access to locally operated models (e.g., Mistral) or optionally to public cloud providers – depending on requirements, costs, and data protection needs of the workload.
Discuss the architectural and product decisions: vLLM as a GPU-optimized inference engine, a central AI gateway that enforces multi-tenancy, security, and comprehensive observability.
A brief live demo will show how applications and developers access AI models through a single API – with token-based quotas per tenant and intelligent routing.
An experience report from an exciting AI infrastructure project in Austria.
-----------------------------------------------------------------------
Networking: --> snacks etc. --> 10:00