A DevOps journey through Alps

For over three decades, the Swiss National Supercomputing Centre (CSCS) has developed and operated Switzerland’s high‑performance computing and data research infrastructure, enabling world‑class science. As scientific workflows have grown more diverse—across both hardware and software dimensions—the complexity of deploying and operating HPC systems has surged, pushing traditional vertically-integrated management models to their limits. Addressing these modern demands required a fundamental rethinking of the entire stack. Alps, CSCS’s newest system, represents the first large‑scale realization: a modular, versatile service‑oriented infrastructure designed to evolve alongside the scientific challenges it supports. This talk shares the journey of such transformation through the lens of an engineering team responsible for core SDLC services. It explores the DevOps challenges and lessons behind this shift and how we continue to adapt along the way.