Speakers
Description
Cross-disciplinary collaboration in academic institutions is frequently constrained by the difficulty of identifying relevant expertise across organizational and disciplinary boundaries. Without effective discovery mechanisms, potential collaborations remain unrealized, limiting the scope and impact of research and innovation. UniFlow is a software platform being developed by Simplera AS in collaboration with UiT The Arctic University of Norway to address this problem. Users describe their project or idea via a guided natural language interface, and the system returns a ranked set of relevant individuals along with their associated fields, publications, and projects, visualized as an interactive network graph depicting the relationships between people, expertise domains, and research outputs. The underlying architecture employs a hybrid retrieval-augmented generation (RAG) approach, combining keyword-based and semantic search over structured research metadata sourced from the Norwegian National Research Archive (NVA). The AI system is built on European large language models and incorporates prompt engineering, output validation, and safeguard mechanisms aligned with GDPR and the EU AI Act. At the time of writing, system design and implementation are actively underway, with structured user testing planned prior to the conference. Although developed in the context of UiT, the platform is designed to be publicly accessible, enabling researchers, students, and external stakeholders to discover and connect with academic expertise.