Lena A Mittmann, Hampus Nässtrom, Eugène Bertin, Sarthak Kapoor, Inês Diogo, Anat Itzhak, Giulia Dalmonte, Rasmus Thorup Danielsen, Malthe Emil Skytte, Holger von Wenckstern, Andrea Crovetto, José A Márquez
High-throughput and autonomous experimentation generate complex, heterogeneous datasets whose reuse, large-scale analysis, and linking to data from first-principles calculations remain challenging. Here, we report on an operational deployment of the open-source NOMAD Oasis data infrastructure as a digital backbone for daily high-throughput experimental work in a materials discovery group at DTU Nanolab. The customized platform supports the complete experimental workflow in a thin-film laboratory, linking physical inventory items, automated synthesis logs, high-throughput characterization data, and cloud-based analysis within a unified data model. The platform preserves full sample provenance across multi-step workflows and uses unified sample coordinates to enable direct cross-technique correlation. We demonstrate the capabilities of the platform on a combined experimental-computational project centered on phosphosulfide thin-film libraries. The data consists of 12 000 experimental samples acquired during routine operation over 18 months, complemented by more than 800 first-principles calculations - all managed in a unified data platform available at https://nomad.nanolab.dtu.dk/nomad-oasis/gui/search/combinatorial-samples. We show how an open, extensible data infrastructure system supports reproducible workflows, scalable analysis, and long-term data reuse.