Welcome!
- zhuohao [at] uw.edu
- Google Scholar
- @ZhuohaoZhang
- Zhuohao (Jerry) Zhang
My name is Zhuohao (Jerry) Zhang. I am a final-year Ph.D. candidate at the University of Washington, working with Prof. Jacob Wobbrock at the ACE Lab. I obtained my M.S. degree in CS from the University of Illinois, Urbana-Champaign, where I worked with Prof. Yang Wang, and my B.Eng. degree in CS from Zhejiang University, China. I also worked closely with Prof. Anhong Guo at UMich.
My research asks how well AI models understand the work people delegate to them and how people can retain authority over that work. I build benchmarks and datasets that measure model judgment like visual design quality and the real-world consequences of agent actions. I use those findings to design human–AI systems that let people make their intent explicit, inspect a model’s interpretation, and verify the result before accepting it. I develop this work with blind and low-vision creators, using accessible visual authoring as a stress test for independent AI verification.
I am an Apple Scholar in AI/ML (2024), one of ~20 PhD students selected worldwide each year, and an Anthropic AI Safety Fellow (2026), selected from more than 10,000 applicants with an acceptance rate below 1%.
Recent Experiences
Selected Research
Slides are usually treated as purely visual. A11yBoard rethinks the screen reader for 2-D canvases, making slides perceivable and editable without sight. It was later deployed for Google Slides (ASSETS 2023).
Can AI critique design reliably? Only with structure: our expert-built taxonomy and 2,400-slide annotated dataset substantially improve models' ability to detect design flaws and to propose fixes that work.
A taxonomy and benchmark for whether AI agents understand the consequences of their UI actions: what is reversible, what is not, and who else is affected. Frontier models still misjudge exactly the actions that carry real risk.
Enables blind programmers to verify AI-generated 3-D models they cannot see by keeping code, semantic hierarchy, AI description, and rendering synchronized, supporting independent modeling that previously required sighted help.
Connects scattered photos, screenshots, and videos into contextualized personal memories so multimodal AI can answer questions spanning events, relationships, preferences, and experiences. It reached 71.5% accuracy and beat or tied a conventional RAG system in 74.5% of comparisons.