I'm a developer who started by writing WeChat mini-programs in 2019 and somehow ended up building AI agents in 2026. The trajectory made sense to me at the time.
Right now I'm deep in the LLM/Agent space — building reusable skill frameworks, interview coaching tools, and anything that makes AI actually useful in the real world rather than just impressive in a demo.
I believe the best prompts are engineered, not guessed. And the best tools are ones you'd actually use yourself.
From mini-programs to multi-agents — one commit at a time.
- AI Career Coaching — an agent that helps engineers structure experience and ace technical interviews
- Prompt Vault — a personal prompt library, because muscle memory is underrated
- Claude Skills — reusable frameworks for interview answers, company research, competitive analysis, bookmark visualization, and more
- Purple Portfolio — a personal website, because every engineer deserves a home on the internet
| Project | What it does |
|---|---|
| 🎯 interview-answer-framework | Transforms any interview question into a structured, impressive answer |
| 🏢 interview-company-research | Deep-dives a company from two angles: how to impress them, and whether they deserve you |
| 🌧️ bookmark-visualizer-skill | Bookmark Visualizer / 及时雨 — Chrome export → CRT terminal nav in one HTML; scanlines, smart buckets, live search |
| 🤖 AI-Career-Coaching-Navigator | Conversational agent for JD-aligned resume generation and interview prep |
| 💜 purple-portfolio-shine | Personal portfolio website — elegant, TypeScript, mine |
| 🗃️ promptvault | Personal prompt library — because good prompts are worth saving |
Languages: TypeScript · Python · JavaScript · HTML/CSS
AI/LLM: Claude · LangChain · AutoGen · Agent frameworks
Past lives: WeChat mini-programs · Django · Vue · NLP · Deep Learning
Currently: Building with Claude Skills, Claude Code, MCP
2019 → WeChat mini-programs, Python basics, falling into NLP
2020 → Deep Learning, sentiment analysis, GANs
2021 → NLP team learning, Datawhale
2024 → AI Agents, AutoGen, LangChain, the rabbit hole opens
2025 → LLM tooling, prompt engineering as a discipline
2026 → Claude skills, AI career tools, this README you're reading
Six years. One direction: make machines do interesting things.
- 🌐 Website: lijiaxing.com.cn
- 🐙 GitHub: @Timelovers
- 📦 Prompt Vault: timelovers.github.io/promptvault
2019 年写微信小程序，2026 年在构建 AI Agent。中间发生了很多事，但方向一直没变过：让机器做有意思的事情。
现在主要在 LLM / Agent 方向深耕——构建可复用的 Skill 框架、AI 面试辅导工具，以及各种「看起来应该存在但还没有人做」的东西。
我相信好的 prompt 是工程出来的，不是猜出来的。
从小程序到多智能体，一个 commit 接着一个 commit。
- AI 职业教练 — 帮助工程师梳理项目经验、精准匹配 JD、搞定技术面试
- 提示词库 — 个人 prompt 知识库，因为好的提示词值得被记住
- Claude Skills — 面试答题框架、公司调研、竞品分析、书签可视化等可复用技能包
- 紫色作品集 — 个人网站，因为每个工程师都应该有个自己的角落
| 项目 | 一句话说明 |
|---|---|
| 🎯 interview-answer-framework | 把任何面试题转化为有结构、有深度的 Impressive 回答 |
| 🏢 interview-company-research | 双视角调研公司：了解它，也评估它值不值得你去 |
| 🌧️ bookmark-visualizer-skill | 及时雨书签可视化 — Chrome 书签导出 → 单文件复古终端导航；扫描线磷光、智能分类、实时搜索 |
| 🤖 AI-Career-Coaching-Navigator | 对话式 Agent，生成匹配 JD 的简历 + 面试故事训练 |
| 💜 purple-portfolio-shine | 个人作品集网站，是我的个人作品集网站，用来展示我是谁、我做了什么、以及我的思考方式。 |
| 🗃️ promptvault | 个人提示词库，好的 prompt 值得被好好存起来 |
主力语言： TypeScript · Python · JavaScript · HTML/CSS
AI/LLM： Claude · LangChain · AutoGen · Agent 框架
前世今生： 微信小程序 · Django · Vue · NLP · 深度学习
现在在用： Claude Skills · Claude Code · MCP
2019 → 微信小程序、Python 入门、一头扎进 NLP
2020 → 深度学习、情感分析、GAN
2021 → NLP 组队学习，Datawhale
2024 → AI Agent、AutoGen、LangChain，兔子洞打开了
2025 → LLM 工具链，把 prompt 当作一门工程来对待
2026 → Claude Skills，AI 职业工具，以及你正在读的这份 README
六年，一个方向：让机器做有意思的事。
「枕风宿雪多年，我与虎谋早餐。」
Slept in wind and snow for years, now having breakfast with tigers.
Made with 💜 by Sara · @Timelovers