AI systems cannot exist without data. Now that AI models (data science and AI) have matured and are readily available to apply in practice, most organizations struggle with the data infrastructure to do so. There is a growing need for data engineers that know how to prepare data for AI systems or that can setup enterprise-wide data architectures for analytical projects. But until now, the data engineering part of AI engineering has not been getting much attention, in favor of discussing the modeling part. In this paper we aim to change this by perform a mapping study on data engineering for AI systems, i.e., AI data engineering. We found 25 relevant papers between January 2019 and June 2023, explaining AI data engineering activities. We identify which life cycle phases are covered, which technical solutions or architectures are proposed and which lessons learned are presented. We end by an overall discussion of the papers with implications for practitioners and researchers. This paper creates an overview of the body of knowledge on data engineering for AI. This overview is useful for practitioners to identify solutions and best practices as well as for researchers to identify gaps.

Background: Software solutions for wearable-based stress monitoring offer significant potential in health care, particularly for vulnerable populations such as individuals with dementia or persistent physical symptoms. Despite technological advances, designing user-centered, ethically grounded, and contextually relevant software remains challenging. Vulnerable populations often have specific cognitive, physical, and emotional needs that require customization, yet these are rarely prioritized in mainstream development. Our so-called Sensors2Care project addressed these challenges by co-developing stress-monitoring prototypes in collaboration with stakeholders from health care, law, and technology within a transdisciplinary setting. Objective: This article has two aims: first, to describe how the Sensors2Care project operationalized the transdisciplinary approach (TDA) within a learning community (LC) to guide the development of stress-monitoring software; second, to share stakeholder needs and design requirements for wearable technologies in complex health care contexts, derived from this process. Methods: The Sensors2Care project applied a TDA embedded in an LC. This approach combined participatory design research with mixed methods across 3 iterative components: requirements gathering, prototype development, and early-stage evaluation. Research activities included scoping reviews, semistructured interviews, focus groups, legal analyses, and field testing. In the LC, students and researchers from health professions, computer science, and law collaborated with patients, (in)formal caregivers, and industry partners in a transdisciplinary consortium. User stories served as both a methodological tool and design outcome, helping to capture stakeholder needs and align input from technical, health, and legal domains. Feedback was collected continuously and used to refine requirements and prototypes throughout the development process. Results: User stories revealed 7 key themes relevant to developing and using wearable-based stress monitoring, including strategic use, notifications, user input, data insight, data access and sharing, hardware design, and support. Stakeholders emphasized the need for customization, durability, and comfort, aligned with the cognitive and physical needs of the target populations. Prototype evaluations indicated the practical relevance of these features and revealed a need for training and insight into long-term usability. Beyond their role in capturing content-driven input, user stories also supported transdisciplinary collaboration by aligning legal, health, technical, and experiential perspectives. This was facilitated by the LC structure, which enabled sustained engagement between students, researchers, and societal stakeholders and illustrated the feasibility of implementing TDA in a university context. Conclusions: This project illustrates how TDA, when embedded in an LC, supports the co-development of ethically grounded, contextually relevant, and practically applicable stress-monitoring software for vulnerable populations. The iterative design process enabled early integration of legal, health, and technical considerations, while user stories supported structured collaboration across domains. Although the project resulted in concrete prototypes and clustered design requirements, further research is needed to assess long-term use and real-world implementation across health care contexts. Embedding TDA in LCs may strengthen future professionals’ ability to address complex health care challenges collaboratively.