Generative AI Engineer in Taichung, Taiwan.
I work on self-hosted LLM infrastructure — vLLM serving, LiteLLM proxy, GPU scheduling on Kubernetes, and the benchmarking to know whether any of it works. Also interested in RAG evaluation and network/infrastructure design.
Most of my production work lives in private repos. What's here is side projects and older work.
kaiyu.io · ORCID · Google Scholar · LinkedIn