About Me

自我介绍

About Me

我是一个爱摆烂的人类。本科(直博)就读于澳大利亚国立大学,计算机专业,是Fabian MuehlboeckStephen Blackburn的学生,主修系统架构。

I’m a human who sometimes likes to lay around. I’m currently an undergraduate (honours) at the Australian National University, majoring in Computer Science, have completed research projects under the supervision of Fabian Muehlboeck and Stephen Blackburn, focusing on systems architecture.

爱好

Hobbies

我的爱好不多,基本就是写写代码,看看番,玩玩游戏。

I don’t have many hobbies—mainly coding, watching anime, and playing games.

编程

Programming

  • 在Unity领域深耕七年,有开源一个开箱即用的热更新框架JEngine
  • 写了八年C#,对该语言的底层优化(和底层架构设计)较为熟悉,推出了以下开源库:
    • 高性能序列化库Nino
    • 高承载多线程TCP网络库Miku
    • 易上手的多线程异步ECS库EasyEcs
  • 在开源社区提交屎山代码六年了,主要 (能记得起的) 对以下开源项目进行了贡献:
    • C#动态解释执行器(C#热更库)ILRuntime
    • 高性能序列化库MemoryPack
    • 应该还有但我记不住了,感兴趣的可以翻我Github主页
  • 最近在研究LLM Agent,打算弄一个写长篇小说的LLM,并且克服LLM模型的一些通病(例如幻觉):Writer
  • Seven years of deep experience in Unity, and I’ve open-sourced a plug-and-play hot update framework: JEngine
  • Eight years of C# experience, with a strong grasp of low-level optimization and architecture. I’ve released several open-source libraries:
    • High-performance binary serialization: Nino
    • High-throughput, multi-threaded TCP networking: Miku
    • Easy-to-use, multi-threaded asynchronous ECS: EasyEcs
  • I’ve been contributing to open source community for six years, and have contributed to these projects:
    • C# dynamic interpreter (hotfix library): ILRuntime
    • High-performance binary serialization: MemoryPack
    • There are plausibly more, but I can’t remember them all, feel free to check my GitHub
  • Recently, I’ve been exploring LLM Agents, aiming to build an LLM for writing long-form novels and tackling common LLM issues (like hallucinations): Writer

动漫

Anime

  • 最近喜欢《青之箱》
  • 《Gosick》也很喜欢
  • 其实看了很多番但在我写自我介绍的时候我基本想不起来了
  • Recently into "Blue Box"
  • Also a big fan of "Gosick"
  • Honestly, I’ve watched a ton of anime, but can never remember them when writing an intro

其他

Others

暂时想不出写什么了,以后再补充吧

Nothing else comes to mind for now—maybe I’ll add more later.

Profiling Source Generator Performance