Over the course of two exciting days, more than 60 students teamed up with experienced mentors from across Europe to explore MLSysOps’ “SysOps in Action”. The MLSysOps framework, built atop state-of-the-art tools such as Karmada, Kubernetes, Grafana, and OpenTelemetry, allowed them to design and apply different policies for cloud-edge continuum computing.

From telemetry collection to cluster-level optimization, participants tackled the complexity of ensuring application performance while reducing operational costs. Besides considering ML models, some teams even experimented with Genetic Algorithms, demonstrating creativity in combination with technical knowledge.

Overall winner: The Red team
Chosen by the jury as the overall winner, considering the approach, platform usage, creativity and presentation quality.

Best presenter: The Purple team
Recognized for the clarity and energy of their presentation.

All participants:
All 7 teams provided valuable insights, demonstrating how diverse perspectives fuel innovation.

Beyond just solutions, the Hackathon left us with stronger skills, fresh ideas, and the kind of energy that only collaboration can spark.

The Hackathon welcomed participants with proactive, creative problem-solving skills who thrive in a fast-paced, collaborative environment. While no specific background was required, those with prior knowledge in Python, REST, ML tools/API as well as a basic notion of concepts such as micro-service paradigm, container, VM, and tools such as Docker and Kubernetes, were able to get the most out of the experience.

In terms of personal equipment, participants merely needed a laptop with a Python environment and ML tool/pipeline to build policies and train/serialize ML models, which they subsequently integrated in the MLSysOps framework via a simple command line interface, to control the way application deployment decisions were made.

September 26, 2025

September 27, 2025