AI 产品经理｜Agentic AI、AIGC、企业智能化与 AI 商业应用
我关注如何把模型能力转化为可靠、可评估、可落地的产品体验，通过清晰的工作流、完整的异常处理和可衡量的评估体系连接技术能力与真实用户任务。
- Agent 产品架构、工作流设计与工具调用
- AI 产品评估：生成质量、任务成功、用户体验与业务价值
- AIGC 与多模态内容生产工作流
- RAG、推荐系统与 AI 辅助决策产品
一本从产品视角出发的 AI Agent 实践手册，包含架构说明、工作流检查清单、评估框架和可复用的产品模板。
一套面向 AI 时代端到端 Ownership 的中英双语方法论，包含真实需求识别、快速验证、使用跟踪、阻碍升级和失败复盘模板。
- 从真实用户任务出发，而不是从模型能力出发。
- 把失败处理、恢复机制和人工介入设计为主流程的一部分。
- 评估完整的产品体验，而不只是单次模型输出。
- 只使用原创内容、公开信息和合成示例进行公开建设。
本主页及关联项目只使用公开信息、原创方法论或合成示例，不包含任何雇主、客户、用户或内部项目的保密信息。
AI Product Manager focused on Agentic AI, AIGC, enterprise automation, and AI-powered commerce.
I turn model capabilities into reliable product experiences through clear workflows, complete failure handling, and measurable evaluation systems that connect technology with real user tasks.
- Agent product architecture, workflow design, and tool use
- AI product evaluation across quality, task success, experience, and business value
- AIGC and multimodal creator workflows
- RAG, recommendation, and AI-assisted decision products
A product-first handbook for designing reliable AI Agent products, including architecture guidance, workflow checklists, evaluation frameworks, and reusable product templates.
A bilingual framework for AI-era end-to-end ownership, with reusable templates for opportunity validation, usage tracking, escalation, and retrospectives.
- Start with a real user task, not a model capability.
- Design failures, recovery, and human review as part of the main workflow.
- Evaluate the complete product experience, not only model output.
- Build in public with original content, public information, and synthetic examples.
Everything shared here is based on public information, original frameworks, or synthetic examples. No confidential employer, client, user, or internal project material is included.