Research data management has become a key topic in modern science. Scientific journals and funding agencies increasingly require that research data be properly stored and made accessible, primarily to support reproducibility. In the Czech Republic, this challenge is being addressed by the National Repository Platform (NRP) project.

Repositories, however, should not only serve as passive storage for downloadable data. They can also provide functionality for browsing, analysing, and visualising stored datasets. In this way, they can evolve into scientific platforms that actively support researchers in working with data.

A robust solution is therefore needed to run workflows that ensure the same set of analytical computations is applied consistently to every deposited dataset. This guarantees that each dataset is accompanied by a uniform set of derived results.

The objective of this thesis is to design and implement a proof-of-concept extension to the repository operated within the National Data Infrastructure. The extension will automatically process uploaded datasets and provide data curators with insight into dataset processing. It must be configurable, modular, and deployable to a Kubernetes cluster alongside the repository. The thesis will describe both functional and non-functional requirements, discuss architectural decisions and technologies employed, and deliver a practical outcome in the form of a component that connects the repository to a batch-processing system together with a Helm chart packaging the repository and the extension’s components.