Computer Engineering undergraduate · HKUST
Multimodal LLMs, AI agents, and mobile AI systems
Hong Kong Hong Kong University of Science and Technology
I am a Computer Engineering undergraduate at the Hong Kong University of Science and Technology (HKUST). I am interested in building useful systems around multimodal large language models, AI agents, and mobile platforms.
My current cumulative grade average is 4.185/4.3, placing me in the top 1% of my cohort. I enjoy moving between research questions, working prototypes, and products that reach real users.
BEng in Computer Engineering · 2024–2028 (Expected Graduation)
Cumulative Grade Average: 4.185/4.3 (Top 1%)
Research experience · arXiv preprint
Research on smartphone-based tour guidance systems using large language models to produce context-aware travel narratives from real-world photo capture.
Research project
Research on routing multimodal queries to improve the cost-performance trade-off of model selection across different input modalities and task complexities.
arXiv preprint · 2026
Automatic photo tour guidance with smartphones and large language models. The project uses everyday mobile capture to provide context-aware tour narration and guidance.
View on arXivFlutter · Android
Independently developed the full-stack Flutter application for critical SMS detection and real-time alerting. Deployed on major Android app stores in China; recorded over 500 active users in the first month and generated initial revenue.
Flutter · LLM application
LLM-powered Flutter application for language learning through customized narrative generation and interactive content design. It embeds target vocabulary into personalized stories to make practice more contextual and engaging.
HK InnoX Summer 2026 · Real-time perception
Built a real-time snow terrain scanning system backed by HongShan's Hong Kong X Foundation. Led algorithm development for real-time terrain perception and snow surface scanning; helped the project win the Best Product Prize and advance into the HKSTP Ideation Programme.
Agent workflows · Study automation
Designed and authored reusable agent skills, including summarize-slides, homework-solver, and resume-crafter, built around multi-agent collaboration for study and resume-writing workflows. Shared publicly, receiving 15,000+ likes and saves on Xiaohongshu and 450+ GitHub stars.
Fall 2024 · Spring 2025 · Fall 2025 · Spring 2026
HKD 40,000