A static site connecting a professional profile, bilingual résumé, engineering projects, and long-term topics.
Content direction and design review, with AI-assisted implementation and validation.
Development · Testing · Inquiry
ELL3N HSU
Python engineer, exploring applied AI.
My work spans backend development and test engineering. I focus on business logic, validation, and maintainable systems, and I am extending that experience into AI workflows and LLM inference.
Experience
Read the résumé → Backend development
Turning business rules into maintainable services and tools.
Test engineering
Translating requirements into verifiable workflows, with testability and automation in mind.
AI practice
Exploring AI-assisted testing workflows and studying inference systems.
Selected work
All projects →A traceable learning record for LLM inference, KV cache, and evaluation methods.
Framing study questions, understanding inference mechanisms, and planning reproducible experiments.
Areas of inquiry
Starting with practice. Going deeper.
Applied AI & inference
From testing workflows to inference fundamentals, KV cache, and reproducible experiments.
Engineering practice
Python, test automation, domain modeling, and maintainable tools.
Recent notes
All notes →Prefill vs Decode: LLM Inference 的两个阶段
理解 LLM 推理中 prefill 和 decode 的区别，以及为什么 prefill 更适合 batching。
KV Cache 为什么会吃显存？
梳理 KV Cache 的数据结构、显存估算方式，以及长上下文为什么会放大问题。
Prefix Cache 命中率如何影响 TTFT？
分析 Prefix Cache 命中与未命中对首 token 延迟的影响，并记录后续 benchmark 计划。
Let's talk engineering and AI.
Find more code and project activity on GitHub.