Research software collaborations encompass various forms such as meetings, informal chats, pull requests, and GitHub issues, which can cause decisions to lose their original rationale. A decision may surface in a Slack thread, be refined in a meeting, and implemented in a pull request.
However, these different modes of communication can lead domain researchers and research software engineers to develop divergent mental models regarding project intent, ownership, and scientific assumptions.
We argue that alignment in research software engineering is a continuous lifecycle problem, and agentic AI can assist in stakeholder alignment and project-state tracking without replacing human decision-making.
We present Aleena, an open-source lifecycle alignment agent that utilizes GitHub as a shared collaboration surface, transforming multi-modal stakeholder interactions into structured project records that surface risks, track open questions, and maintain decision continuity.
Grounded in experiences from university-based research software engineering centers, this paper outlines the motivating problem, system design, prototype, and illustrative lifecycle scenarios for Aleena.
Blogger's Review: Aleena emphasizes the importance of maintaining decision consistency in multi-party collaborations. By converting diverse forms of communication into structured records, it effectively reduces misunderstandings and discrepancies. This approach has broad applicability in the field of research software engineering, particularly in facilitating interdisciplinary collaboration. Its open-source nature also provides a solid foundation for future expansions and improvements.