Expert Discovery Platform: RABBIT
Led the development of an AI-powered semantic search platform that helps UW–Madison researchers, industry partners, and campus collaborators discover faculty expertise.
Overview
RABBIT (Research and Business-Bridging Intelligence Tool) replaces keyword-based faculty search with semantic matching, enabling more accurate and comprehensive discovery of expertise across UW–Madison.
Problem
At a large research institution, identifying faculty expertise often relied on manual Google searches and individual profile pages—an approach that was inefficient, incomplete, and limited in surfacing multidisciplinary connections.
Approach
Designed and implemented a semantic search system that matches user queries to faculty scholarship. Results are re-ranked based on multiple signals—including relevance, recency, and research impact—to provide more meaningful recommendations.
Constraints
- Integrate with existing faculty data across multiple internal and external systems
- Restrict access to faculty, staff, and WARF employees via NetID authentication
- Surface relevant, contextual information to support rapid research-to-expert matching
Key Decisions
Custom-built platform instead of off-the-shelf marketplace solutions
Commercial tools did not support the institution’s specific data sources, authentication model, and semantic matching requirements.
Semantic search over keyword matching
Grant proposals and industry challenges are written in natural language. Keyword search often misses relevant faculty using different terminology to describe similar concepts.
Result & Impact
Accelerates faculty discovery for industry partnerships and federal grant applications, particularly those requiring interdisciplinary and cross-sector collaboration.
Learnings
- Close collaboration between technical teams and business engagement stakeholders is critical to building tools that see sustained adoption.
- There is significant demand for structured, institution-wide expert discovery in academia.