Hi, I'm Ray, a final year PhD student at the University of Washington. I design, build, and evaluate human-centered AI systems that support decision-making and human collaboration. I'm also a registered dietitian with clinical experience in mental health and industry experience in health tech.
I am a PhD candidate studying human-centered AI at the University of Washington, advised by Ari Pollack, Wanda Pratt, and Xuhai “Orson” Xu. My research combines human-centered design, AI/ML, and mixed-methods evaluation to study how agentic AI systems support human collaboration and decision-making in multi-stakeholder environments. I am especially motivated by building technologies that measurably improve people’s lived experiences and developing frameworks that help researchers and practitioners build more responsible, trustworthy AI. My research has appeared in high-impact venues in human–computer interaction (HCI) and health informatics such as ACM CHI conference, AMIA conference, and Nutrients journal.
Before my doctoral studies, I completed my master’s degree at the University of Michigan and then worked as a dietitian and research scientist across health organizations and tech startups such as Impossible Foods, Unita Health, and Dexcom.
I’m inspired by the late, great Kobe Bryant’s Mamba Mentality — his relentless commitment to inspiring those around him. If my work resonates with you, I’d love to connect and chat. Outside of research, I enjoy traveling and playing sports, mainly basketball and golf.
Go Blue!Go Dawgs!
Open to Opportunities
Currently seeking internship and full-time opportunities in the following areas for Spring 2027 and beyond!
Generative AIHuman-AI CollaborationResponsible AIAgentic AIHealth & Social Technologies
🔭 Current Work
User Trust towards AI Features
Human-AI CollaborationResponsible AI
I developed an interactive QA system using React that presents users with AI-generated responses augmented by trust-enhancing cues and measures whether those signals promote appropriate trust. Findings will inform design guidelines for trustworthy AI interactions that support appropriate rather than blind trust.
Building AI Digital Twins for Human-AI Collaboration
Agentic AIHuman-AI Collaboration
Health consumers increasingly use AI tools in isolation from healthcare providers, generating risks of hallucination, erosion of clinician authority, and fragmented decision-making. We are investigating what it takes to construct accurate digital proxies—AI agents that can represent individuals to support human-AI collaboration in this setting.
AI-Assisted Values Elicitation
Human-AI CollaborationResponsible AI
My team developed a React application that uses LLMs to elicit, extract, and visualize individual values from open-ended conversation, mapping them onto a shared representational space. This work explores the capacity of current LLMs to act as value mediators, bridging personal values and goals between humans to support collaborative decisions.
📚 Selected Publications
1.ExTENDS: Empowering Domain Experts to Build Personalized AI Proxies for Asynchronous Patient Care
Agentic AIHuman-AI Collaboration
Ray-Yuan Chung, Muying Li, Millie Wu, Ari Pollack, Wanda Pratt, Lena Mamykina, Xuhai "Orson" Xu. In Submission (2026).
2.“It Helped Me Understand, But…”: Examining Trust Calibration in Health AI Among Lay Health Consumers and Patients
Human-AI CollaborationResponsible AI
Ray-Yuan Chung, Athena Ortega, Xuhai "Orson" Xu, Wanda Pratt, Ari Pollack. In Submission (2026).
