I build AI that actually gets used — starting from real problems, not benchmark leaderboards.
I'm a CS graduate student at NYCU. I'm drawn to AI that reaches beyond the lab — systems people can actually use to solve real problems. Whether it's sports analytics, open-source generative tools, or image restoration, I'm most motivated when there's a concrete use case on the other side of the work.
My coursework in compiler design and parallel computing gave me a systems foundation I lean on constantly: understanding failure modes before designing solutions, thinking about latency alongside research metrics, and caring about what it takes to go from a working experiment to something reliable in practice.
Where the thinking started.
Vision-language models, agentic system applications, real-time embedded systems, and parallel programming
Focused on computer science fundamentals
What I've actually built, organised by the skills it took to build it.
Notes on reproducibility, tooling, and the small things I keep relearning.
Best by email. I read everything; replies usually within a week.