The Bundesdruckerei is widely recognized for printing German currency and securing patent documents, but it also plays a significant role in government technology services.

Beyond its legacy in secure printing, it provides cloud and platform solutions for various public-sector entities.

PLAIN stands for Platform Analysis and Information System, an ambitious government-wide effort to unify data sharing and analytics under one secure, scalable umbrella.

The initiative was born from a mandate requiring each government branch to stand up its own data science department and collaborate more efficiently on joint projects.

Since many branches lacked both the infrastructure and the expertise to kick-start such initiatives, the Bundesdruckerei took on the challenge of building a platform that offers IaaS, PaaS, and SaaS capabilities. A dedicated data center was erected specifically for PLAIN, ensuring robust security and compliance from the ground up.

Second, I took part in building specific AI applications for various government branches, ensuring that each data science department could hit the ground running with ready-made solutions.

The result was an “app store” model, where data science teams could share their work, collaborate on experiments, and even combine datasets if properly authorized.

Over time, I assembled and led multiple AI app development teams, pairing data scientists with software engineers to bring these applications to production-ready status.

This approach not only accelerated the roll-out of new tools but also fostered a stronger sense of interdepartmental cooperation. Through a mix of secure networking, integrated MLOps practices, and user-friendly deployment workflows, PLAIN quickly became the backbone of data innovation in the public sector.

Ultimately, it stands as a testament to how centralized infrastructure, carefully designed governance policies, and user-centric development can empower a diverse array of governmental organizations.

Initially, a handful of engineers (5-10) kicked off the project, focusing on the fundamentals—network configuration, baseline security, and minimal tooling.

Over time, the success of the pilot phase led to rapid scale-up, and the project team ballooned to around 50-80 skilled professionals. I was heavily involved in two main efforts: fortifying the DevOps-driven PaaS layer and shaping the AI application ecosystem.

On the DevOps side, I helped devise processes for automating analytics workspaces, implementing robust MLOps pipelines, and setting up data flows that could be shared across branches under strict security agreements. This made continuous model training, testing, and deployment a reality for multiple teams at once..

Keywords:

Management, DevOps, MLOps, Platform, AI Apps, Data Mangement, Data Analytics

Previous

Federal Foreign Office
Next

Henkel