High-performance computing (HPC) is a multidisciplinary field where hardware specialists, domain scientists, data stewards, and software engineers work in concert to accelerate research and innovation through large-scale computing capabilities. In recent years, federated data spaces, FAIR data management, software provisioning, and AI and quantum computing services have emerged as core growth areas that are reshaping how supercomputing facilities serve their scientific and industrial communities. A parallel transformation is underway across academic institutions, where research software and research data are increasingly recognised as cornerstones of scientific innovation rather than secondary outputs of the research process.
Many academic institutions have established Research Data Management (RDM) units to assist domain scientists in creating, analysing, sharing, and archiving scientific results in line with the FAIR principles [1]. These guidelines safeguard the reproducibility and reusability of the data, and are now a baseline expectation for many funding agencies and scientific journals. The same requirements apply to the software and workflows used to generate scientific results. In response, a dedicated professional role has emerged: the Research Software Engineer (RSE). RSEs assist domain scientists in writing code aligned with the FAIR principles for research software [2], in the initial porting of serial applications to parallel programming models, and in growing user communities through training events, collaborative coding, and community outreach. Behind these efforts stands a community that has professionalised in recent years, developing its own identity [3] and organising dedicated events, some held independently, others co-located with HPC [4][5] and software engineering [6][7] conferences to strengthen cross-community collaborations.
RSEs typically start out as domain scientists with software development experience, and quickly acquire specialist skills that make them instrumental to the long-term sustainability of the software they maintain. They are usually employed directly by research groups, in close contact with developers and users, and have detailed knowledge of the scientific research enabled by these codes. However, not every lab has a need for a full-time RSE or can afford one over an extended period of time. The position paper "Establishing central Research Software Engineering units in German research institutions" makes the case for establishing central RSE units within research institutions, applying the same pooling logic that already underpins central HPC and RDM services: consolidating institutional expertise, mutualising costs, and keeping specialist skills physically close to the domain scientists they serve. The proposed model is complementary to, rather than competitive with, existing HPC organisations. Supercomputing facilities remain the primary hub for specialised training and software porting workshops, while RSE units catch the long tail of domain scientists who require mentoring before they are ready to engage with HPC at scale.
[1]: Wilkinson et al. "The FAIR Guiding Principles for scientific data management and stewardship". *Scientific Data* 3, 160018 (2016). doi:10.1038/sdata.2016.18
[2]: Barker et al. "Introducing the FAIR Principles for research software". *Scientific Data* 9, 622 (2022). doi: 10.1038/s41597-022-01710-x
[3]: Dehne et al. "Research Software Engineering for Natural Sciences". *Softwaretechnik-Trends* 45 (2025). https://dl.gi.de/handle/20.500.12116/47200
[4]: RSE HPC conference co-located at ISC24: https://www.rse-hpc.org
[5]: RSE-HPC-2025 conference co-located at SC25: https://us-rse.org/rse-hpc-2025/
[6]: Software Engineering and Research Software conference co-located at ICSE 2026: https://conf.researchr.org/home/icse-2026/sers-2026
[7]: deRSE25 conference co-located at SE25: https://events.hifis.net/event/1741/overview