Technical thinking starts here 这里将会分享我的技术思考与见解

This is the first post on the blog.

I want this place to hold technical thinking that is closer to field notes than tutorials: what broke, what I tried, what patterns started to show up, and which assumptions did not survive contact with the work.

The writing will probably orbit around a few themes:

• AI systems and agent engineering: reliable LLM products, reasoning workflows, retrieval, and evaluation
• Software architecture and engineering practice: trade-offs, code structure, and maintainability over time
• Tools and workflows: making the development process itself faster, clearer, and easier to inspect

Most entries will stay short. They do not need to be complete to be useful. Each post should still capture one honest unit of thought.

If those topics sound relevant to you, check back soon.

这是这个博客的第一篇文章。

我打算在这里持续记录我的技术见解, 但会更像工程现场笔记, 而不是教程或文档: 遇到了什么问题, 尝试了什么方案, 发现了什么规律, 以及哪些假设被推翻了。

涉及的方向可能包括：

• AI 系统与 Agent 工程：构建可靠的 LLM 应用，推理链设计，检索增强与评测方法
• 软件架构与工程实践：系统设计的取舍、代码组织以及长期可维护性
• 工具与工作流：如何让开发过程本身更高效、更清晰、更可观测

内容会以短文和笔记的形式出现。它们不一定完整，但每篇都会是一个真实、独立的思考单元。

如果你对这些话题感兴趣，欢迎回来看看。