Product manager, former product designer, and independent builder.
I work on products where AI, complex workflows, and human judgment meet. My background spans B2B platforms, cloud security, and regulated digital-asset infrastructure. More recently, I’ve been building products end to end and documenting what I learn along the way.
An AI career workbench connecting job discovery, evidence-grounded fit analysis, application preparation, and tracking.
An AI feedback-triage system that turns unstructured feedback into traceable work packs, with source grounding, evaluation, deterministic checks, and human review.
Live demo · Source · Case study
A practice companion for adult piano learners, designed to make limited practice time more focused and useful.
A privacy-first tool for organizing important personal information into a clear handover for the people who may one day need it.
A small set of field notes from building AI products — focused on product architecture, human/model boundaries, reliability, cost, and AI-assisted delivery.
- AI Cost Is a Product Architecture Problem — Why model spend becomes a product and systems decision once AI is embedded in a real workflow. 中文
- Model + Harness + Product: What Makes an AI Product Work — Why real AI product capability depends on context, tools, runtime, permissions, verification, and user control — not only the model. 中文
- What Product Managers Should—and Shouldn’t—Delegate to AI — AI can expand options and expose gaps; consequential product choices still need explicit human ownership. 中文
- From Linear Handoffs to Spiral Building with AI — How AI-assisted building changes when implementation moves faster than product judgment, and why deliberate product, design, and engineering checkpoints still matter. 中文
100 Days Building — a public experiment documenting what I learn while building independent products, primarily in Chinese on RedNote.
RedNote: Limin Zheng