FDZ UB Mannheim: Three challenges in establishing sustainable research data management, data science, and AI services

FDZ UB Mannheim: Three challenges in establishing sustainable research data management, data science, and AI services Dr. Renat Shigapov, Dr. Irene Schumm, Jan Kamlah,Larissa Will, & Thomas Schmidt 16.09.2025, Data-Science-Community Workshop 2025 Scope Partnerships Sustainability Agenda 1. Introduction 2. Define scope 3. Build partnerships 4. Plan sustainability 5. Conclusions 2 Emerging data science & AI services •Data science services rapidly emerge at universities worldwide •International examples: Harvard, Utah, Purdue, NC State, and Groningen •National example: Data Science Center at the University of Bremen •NFDI: NFDI4DataScience 3 All We FDZ UB Mannheim aligns with international trends Question & Challenges 4 How do we establish RDM, data science, and AI services sustainably? Scope Partnerships Sustainability Herndon, J. (ed.): Data Science in the Library: Tools and Strategies for Supporting Data-Driven Research and Instruction. Facet, 2021. https://doi.org/10.29085/9781783304615 “All libraries face three main challenges in developing a new service.” Agenda 5 1. Introduction 2. Define scope 3. Build partnerships 4. Plan sustainability 5. Conclusions Our key idea: Scope is not fixed – it evolves with demand, feasibility, and collaborations •Dynamic & evolving portfolio –RDM, Data Science, and AI services –Scope defined flexibly, guided by researcher needs •Data-driven approach –Feedback from consultations, trainings, and outreach –Usage statistics & consultancy documentation –Pilot → test promising topics before formal adoption •Learning-by-doing –Team explores emerging trends (e.g., generative AI, LLMs, and RAG) –Prototype solutions → training materials & services 6 Our evolving scope 1. Data-Science-as-a-Service –OCR, transcription, web scraping, APIs –DS pipelines, LLM fine-tuning, and chatbots 2. Education & Training –Courses & hands-on workshops on AI & RDM –https://zenodo.org/communities/fdzmannheim 3. Consultancy –Channels: emails, meetings, workshops –Regular hours (OCR, RDM & knowledge graphs) 4. Community Engagement –Mannheim Center for Data Science (MCDS) –GESS Research Day & university outreach 7 5. Data Offers –Licensed databases + in-house datasets –Examples: Aktienführer archive, Reichsanzeiger, Bankruptcy DB, Wer Leitet KG 6. Software Offers –GitHub repos at https://github.com/UBMannheim –AI chatbots for RDM: FAIR GPT & FAIRplexica 7. RDM Resources –Homepage + „Awesome RDM“ GitHub Repo 8. Secure Environment –Data Cube for privacy-compliant sensitive data analysis 9. Research Data Infrastructures Agenda 8 1. Introduction 2. Define scope 3. Build partnerships 4. Plan sustainability 5. Conclusions We build multi-level collaborations 1. Internal (Library) –UB IT → infrastructure support –Subject librarians → promote services 2. University-level –Committees → policy & strategy alignment –MCDS → “academic speed-dating” events –ZLL → events on AI in higher education –Uni-IT → more GPUs –Researchers using our services 9 Help us to improve the RDM, DS, and AI Services at the University of Mannheim! 💡 Do you have ideas? 🤝 Interested in collaboration? 📩 Get in touch 16 Dr. Renat Shigapov: [email protected] Thomas Schmidt: [email protected] Dr. Irene Schumm: [email protected] Jan Kamlah: [email protected] Larissa Will: [email protected] FDZ: [email protected] https://fdz.bib.uni-mannheim.de