With the Semantic Co-Working Space (SCS), the joint project SODa (Sammlungen, Objekte, Datenkompetenzen - Collections, Objects and Data Competence) is establishing a modular, scalable infrastructure for research, teaching and skills development in the field of research data management (RDM) and data-driven research. The focus is on digital work with collection data, which is supported by WissKI (Wissenschaftliche Kommunikationsinfrastruktur - Scientific Communication Infrastructure), OpenRefine, Jupyter Notebooks, WebProtegé and various database systems. The technical infrastructure is operated on a long-term basis by the IGSD e.V. (Interessengemeinschaft für Semantische Datenverarbeitung e.V.) and the Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU). SODa combines a unique constellation of collection practice, humanities research, didactic and IT expertise. In addition to the technical infrastructure, a wide range of digital teaching and learning resources are being developed, including interactive online courses, tutorials in Markdown/LiaScript, project-specific instructions and documentation as well as community-based support via issue trackers and chat channels. These offerings are systematically evaluated in terms of completeness, comprehensibility and user-friendliness and expanded into modular learning units, including Open Educational Resources (OER). In addition, seven experts are available for individual support and consulting. The aim is not only to teach technical skills, but also to promote a structured understanding of the entire research data lifecycle. Real data from collections is used for this purpose, which introduces users to semantic data management in a collaborative, hands-on setting. The integration of a didmos-based system for Identity & Access Management (IAM) as part of the IAM4NFDI Incubator helps to create centralized and secure access to the services. As part of the SODa project, a SWOT analysis was carried out on the existing WissKI documentation. The measures derived from this are aimed at target group-specific content preparation and didacticization, and a consistent focus on specific user needs mapped to relevant learning objectives. In addition to traditional end users, data modelers, developers, and system administrators are also addressed in order to enable the sustainable promotion of digital skills across all project roles. With the SCS and the accompanying training courses, SODa makes a contribution to the RDM and data science community in two ways: Firstly, a robust, semantically sound infrastructure is provided that enables the context-rich collection and use of research data in line with the FAIR principles and ensures the sustainable storage of collection data. Secondly, educational programs will be implemented to provide users with concrete skills in dealing with digital tools, standards (e.g. CIDOC-CRM), data quality and sustainable storage. These offerings are continuously developed and communicated by the existing technical experts who provide advisory support. The integration into the NFDI (including NFDI4Culture, NFDI4Objects, NFDI4Memory) is currently being prepared technically and organizationally. The SCS is thus not only a practical and innovative space for digital collection research, but also a sustainable model for the promotion of RDM, data literacy and competence, infrastructure development, and data science in the field of scientific university collections.