When Engineering Meets Care Coordination: A Student's Journey to Solve Health Care's Invisible Bottleneck

Health care referral coordination remains one of the most persistent inefficiencies in the U.S. health care system, with millions of annual referrals still managed through phone calls, faxes, and manual spreadsheets. This article documents one computer science Ph.D. student's journey from personal frustration to building Carenector, an AI-powered care coordination platform. Drawing on knowledge graph modeling, privacy-preserving ephemeral computation, and simultaneous constraint satisfaction, the platform reframes the referral process as an engineering matching problem. The system's knowledge graph represents facilities and services across multiple regions, updating dynamically to reflect real-time availability, without storing any personally identifiable patient information. Initially deployed as a consumer-facing tool for individuals and families, the platform has since expanded to an institutional version targeting hospital discharge planning teams, with pilot results demonstrating significant time savings per referral. This work highlights how computational engineering principles, applied to workflow design, data modeling, and human-AI collaboration, can transform even the most entrenched manual processes in health care.