AI Practice Builder · AI 实践建造者
Building practical AI for everyday work & life.

探索 AI 如何成为非技术人的工作与生活方法。

AI 实践建造者，现居中国杭州，在 Ant Group 从事产品运营。

我致力于探索 AI 如何成为非技术人的工作与生活方法：AI如何帮助人增强判断、推动执行，并把有效经验沉淀为可复用的工具与系统。

工作之外，我养了一只猫和一只狗，喜欢游泳和桨板等运动，也喜欢养植物。同时也在持续创造能为生活带来微小而美好的便利小产品。

I'm an AI Practice Builder based in Hangzhou, China, working in Product Operations at Ant Group.

I explore how AI can become a practical method for non-technical people—helping us think more clearly, move work forward, and turn effective practices into reusable tools and systems.

Outside work, I live with a cat and a dog, enjoy swimming and paddleboarding, and love caring for plants. I also build small products that bring a little more ease and delight to everyday life.

如果你对我的开源项目有任何疑问，欢迎交流沟通，希望和更多“骄傲工作家”成为好朋友 Questions about my open-source projects are always welcome: soyoungzsy@foxmail.com.

• AI for non-technical work — practical workflows for product operations, research, communication, and everyday decisions.
• Human–AI collaboration — task routing, evidence boundaries, waiting states, and verifiable completion, reflected in AI Co-brain Workflow.
• Knowledge that improves with use — local-first systems that learn from decisions and feedback, reflected in Self-enhancing Knowledge System.
• Small tools that change a real habit — from immersive English reading with SOYA LingoLens to reflective decision-making with SOYA Personal Board.

• AI-native ways of working — route work between humans and AI, keep evidence visible, and define what “done” actually means.
• Self-enhancing knowledge systems — turn notes, decisions, and feedback into a local-first system that keeps getting better.
• Small products with real utility — build focused tools for learning, reflection, and everyday life.

Log better inputs → Orient with evidence → Operate toward a verifiable result → Preserve what can compound.

I call this the SOYA AI LOOP: a practical way to make AI part of the work itself, while keeping human judgment in charge.

Evidence before claims · Judgment before automation · Local-first by default · Useful enough to become a habit

I share tools and methods that have survived real work—not just polished demos.

Build with judgment. Finish with evidence. Preserve what compounds.