English · 中文

I build AI product prototypes around real workflows: helping retail staff create grounded content, making ESG information easier to explore, and turning complex work into clear next steps.

I'm a 2027 graduate at Zhejiang University, exploring opportunities in AI Product Management. My work connects problem framing, product decisions, hands-on prototyping, and evaluation.

An AI marketing workbench for Huawei PC retail scenarios. Start with an audience or an idea, select a product, create platform-specific copy and images, check product references, and prepare customer follow-up drafts.

Inside the case: conversational clarification · platform Skills · image sub-agent · evidence tracing · price comparison · human review.

A consumer-facing ESG search prototype. Search a brand, explore AI-analyzed environmental, social, and governance signals, and follow the source links behind the key facts.

Inside the case: search-to-analysis workflow · explainable score composition · source-linked facts · scoring limitations.

A product case for manufacturing inspections: turn regulation text into reviewable check-item drafts, carry them into field inspections, and track issues through remediation and verification.

Inside the case: regulation-to-check Skill · citation checks · applicability conditions · human approval · workflow states.

A university–industry matching prototype. Help companies describe technical needs, understand research capabilities, and move from a shortlist to a structured cooperation discussion.

Inside the case: two-sided workflows · rule-based requirement analysis · matching explanations · public research evidence.

• Start with the task. Define who needs help, what they are trying to finish, and where the workflow breaks.
• Build the smallest useful loop. Make the core journey tangible before expanding the feature list.
• Make AI inspectable. Show evidence, surface missing context, and design where a person takes over.
• Separate demos from impact. Working prototypes are evidence of implementation; adoption and business outcomes need their own validation.

These repositories include prototypes and product case studies at different stages. Each project explains what is implemented, what is illustrative, and what remains to be validated.