Favicon Logo

📣️ New: CRAN R package binaries for arm64 and Alpine Linux

devXY's Scope

devXY was established with the goal of promoting the seamless integration of DevOps and Data Science at a professional level. In recent years, projects that blend both disciplines have become increasingly common. Effective communication and alignment of goals among all stakeholders are critical for implementing efficient workflows and ensuring successful collaborations.

The mission of devXY encompasses the configuration, optimization, and maintenance of environments that support both Data Science and DevOps, delivering value to users while remaining easily maintainable through semi-automated installation processes. This approach is intrinsically linked to containerized workloads—whether in CI/CD pipelines or Kubernetes environments—and involves the creation and management of related assets such as packages, images, and charts. devXY leverages these components to enhance existing standards and support projects that bridge Data Science and DevOps, whether by hosting Shiny Apps in R, developing APIs to train machine learning models on GPUs using Python, or managing custom container images.

Personal Background

Over a decade ago, during my time at university, I was first introduced to Data Science. Beyond building models, generating summary tables, and creating boxplots, I quickly realized that completing projects required much more: repeating experiments, publishing results, setting up isolated environments, and regularly updating the toolset. At the time, there was no established infrastructure to support these tasks, so I took the initiative to build it myself. I installed IDEs, configured job schedulers, set up operating systems, and created containers. It became clear to me that of these elements were essential for the success and efficiency of any Data Science endeavor.

During my PhD studies (supervised by Prof. Alexander Brenning), I took on the responsibility of planning, installing, and configuring a 6-node SLURM cluster for the university’s department, designed to train Machine Learning models. This cluster remains in active use today, providing invaluable support for ongoing academic research projects. Additionally, I contributed as a core developer to the mlr3 machine learning framework in R, further expanding my expertise in the field.

Today, after years of combining expertise in both fields while leading the infrastructure/DevOps team at cynkra GmbH from 2020-2024, and deepening my focus on automation and deployment, I am eager to share my knowledge with others. Having contributed to numerous FOSS projects, I remain actively involved in the development and maintenance of public infrastructure components. This includes work on the official Gitea and Woodpecker CI Helm charts as well as the Ansible-based infrastructure setup for Codeberg. These contributions allow me to continuously get in touch with new developments and engage with the FOSS community.

Personal Picture - 3

Name Origin and Pronouncation

The name devXY is derived from two components: “dev” representing “DevOps” and “XY” symbolizing the axes of a plot, which refers to the “Data Science” aspect. It is pronounced “dev” + “eks” + “why” = “dev-eks-why,” phonetically transcribed as /dɛv ɛks waɪ/.

Infrastructure and Tools

devXY operates its infrastructure on a bare-metal k3s cluster hosted on Hetzner Cloud, utilizing the terraform-hcloud-kube-hetzner Terraform/OpenTofu project. The cluster applications are managed through Argo CD, with storage, including RWX, provisioned via Longhorn. For network load balancing, the cluster employs Klipper, while internal networking is secured through Wireguard encryption. The mixed-architecture nodes are powered by Micro OS.

The website has been developed using Astro without relying on any JavaScript framework (on purpose), making it blazingly fast.