The current decommissioning of nuclear power plants in Germany is a complex and long-term process that requires specialist staff. However, appropriate programs to qualify such personnel have declined massively since the nuclear phaseout in 2023, plus a substantial number of employees will retire in the next few years (Kettler et al., 2025). To nevertheless keep training decommissioning experts, we consider developing and establishing a domain-specific community platform a promising approach where future staff can acquire competencies through web-based formats and by connecting and collaborating with peers and, in particular, experts. A community platform refers to a social learning platform for a (learning) community of practice (e.g., Pyrko et al., 2019) that integrates features for news, communication, and collaboration with (references to) educational programs and resources as well as directories of institutions, experts, or projects. However, few such complex platforms (e.g., LinkedIn Learning) exist, and sound domain-specific design principles are missing, in particular for the nuclear power sector. In this regard, established models like the Technology Acceptance Model (TAM; Davis, 1989; Venkatesh & Bala, 2008) suggest that users are likely to accept and use emerging technologies when considering them highly useful and easy to use. Consequently, we aim to investigate how a community platform (for the nuclear power sector) can be designed to maximize perceived usefulness, user experience, and, consequently, user acceptance. To tackle this issue and ensure user-centered design, we carried out two studies. Addressing user-friendliness prerequisites, we first performed qualitative shadowing (McDonald, 2005) by observing user experience experts and novices (N = 5) participating in task-based user trials with a comparable platform and subsequently conducting in-depth interviews. Findings, in particular, show that the seamless integration of the platform areas, an intuitive search function, and the free accessibility of all basic functions contribute to high ease of use. To explore usefulness requirements, we conducted a standardized survey, including questions from the TAM questionnaire (Davis, 1989) and opened-ended questions, to assess future users’ perceived usefulness and use intention of potential platform features. In short, preliminary results of the currently still ongoing survey indicate that potential users (N = 38), including nuclear power scientists, project engineers, and managers, find rather conventional educational resources like specialist publications, learning materials, and directories of training courses, particularly useful while being more reserved towards social features like a personalized feed, an event management system, or a messaging service. Results provide fruitful design recommendations for the planned nuclear power community platform and indicative guidelines for designing and optimizing comparable platforms.