algo@rondo:~$ whoami
Algorithm Engineer @ RTK industry
algo@rondo:~$ bias
robustness > hype && accuracy > decoration
algo@rondo:~$ side-channel
MusicI build algorithms that are supposed to keep working after the demo ends.
My day job lives in the RTK world: SLAM, point cloud processing, integrated navigation, and a lot of C++.
Lately, a growing part of my attention has shifted toward LLM systems and the engineering patterns around them.
我做的不是只在演示里好看的算法，而是要在真实环境里继续工作的算法。
我的主要工作在 RTK 相关场景里，核心方向是 SLAM、点云处理、组合导航，以及大量基于 C++ 的算法实现。
最近我也越来越关注 LLM 系统，以及它背后的工程方法和系统设计。
AlgoRondo comes from two things I like: algorithms and music.
A rondo returns to a theme, but never in exactly the same way twice.
That feels familiar: structure, recurrence, variation, timing.
AlgoRondo 这个名字来自我喜欢的两样东西：算法 和 音乐。
Rondo 是一种不断回到主题、但每次又带着变化的结构。
这和工程很像：结构、重复、变化、节奏。
| Domain | Notes |
|---|---|
C++ |
My default language for algorithm work |
LiDAR SLAM |
Geometry-heavy systems, estimation, mapping, robustness |
Point Cloud |
Processing, representation, filtering, and spatial reasoning |
Integrated Navigation |
Multi-source fusion, system behavior, stability |
Classical CV |
Traditional computer vision methods are still part of my toolbox |
LLM |
Current curiosity and likely next growth vector |
-
robustness > hype -
accuracy > decoration -
explainability > mystery -
systems that survive noise, drift, and messy data -
鲁棒性比热闹更重要 -
准确性比表面效果更重要 -
可解释性比神秘感更重要 -
我更相信那些能扛住噪声、漂移和脏数据的系统
Most of the substantial engineering work I do lives in private company repositories, so I cannot show those projects directly here. This profile is not a public dump of everything I have built. It is a compact map of what I work on, how I think, and where I may be heading next.
我很多核心工程工作都在公司的私有仓库里，因此没法直接把项目公开展示在这里。 所以这个主页不是一个“全部作品公开陈列区”。 它更像是一张压缩过的技术地图，说明我在做什么、怎么思考，以及接下来可能走向哪里。
-
Still grounded in geometry-driven systems
-
Looking more seriously at
LLMapplications and evaluation -
Curious about where robotics pipelines and foundation models may eventually meet
-
Usually thinking with music on
-
仍然扎根在以几何和估计为核心的系统里
-
正在更认真地理解
LLM应用与评估 -
对机器人算法链路和基础模型未来的交汇点很感兴趣
-
大多数难题，都是戴着耳机想出来的