Experience
AI & Cloud Solutions Engineering Intern and Project Team LeadChina Mobile Group Guangdong Co., Ltd. / Department of Science and Technology Innovation
I wanted to understand what happens when an AI system leaves a laptop and has to work through hardware, services and a visitor's attention. The exhibition-guide project gave me that full chain to build with He Yutong.
Led a two-person team developing a Rokid AR exhibition guide across smart glasses, an Android companion app and a Python backend, connecting visual recognition with exhibit explanations. I also evaluated agent search, tool use and scheduled tasks, and worked on task feedback and exception alerts.
The aim was a guide visitors could follow without constantly looking away from the exhibit. Alongside development, I prepared a talk on agent-memory reliability and learned to explain technical choices beyond the implementation team.
Development, demonstrations and evaluation informed the handoff; further field work is a separate stage.
Open case study ↗Honours BSc / Computer ScienceUniversity of Toronto, St. George
I'm studying how software is designed, how algorithms are reasoned about and how data is modelled. My current coursework brings together Java software design, computational reasoning, calculus and probability, alongside a foundation in linear algebra.
Outside class, AllAbout Campus connects those ideas to a familiar student problem: finding the right course information. Memory Guard explores how an assistant should respond when its stored context is uncertain.
Open case study ↗Volunteer AI teachingRural students in Xinshao, China
I wanted students with fewer opportunities to encounter computing to have a starting point for understanding AI. The lessons covered computers, classification and simple neural networks.
The classroom had no computer for the slides I'd prepared. I changed the explanation to a Peppa Pig robot story. It became an exercise in making a technical idea understandable with the resources already in the room.
For me, the social purpose was access: make the first conversation about AI possible without requiring a well-equipped classroom.
Data analysis & machine learningJohns Hopkins summer program
Before university, I used Python and XGBoost to study obesity risk factors in a Johns Hopkins summer program. It was an early opportunity to connect machine learning with a question about public health.
The project was educational analysis, not a clinical tool. Its place in my portfolio is the beginning of an interest in using data to understand problems that matter beyond the screen.