AI products × real-world operationsAI 产品 × 真实业务
Düsseldorf, DE
I’m Tong Wu. I connect AI engineering, complex project delivery and product growth—with hands-on experience in European logistics and supply chains.我是吴桐。我把 AI 工程、复杂项目交付与产品增长连接起来，也有欧洲物流与供应链的一线经验。
Current career focus当前求职方向
Complex project management · Product growth · FDE复杂项目管理 · 产品增长 · FDE
AI Automation & Product Manager · Novada LabsAI 自动化与产品经理 · Novada Labs
中文 · English · DeutschNative · C2 · C1母语 · C2 · C1
EU driving licence欧盟驾照EU Blue Card · Based in Germany欧盟蓝卡 · 常驻德国
Economics → supply chains → AI products经济金融 → 供应链 → AI 产品
02 /Agent memory, orchestration & working products智能体记忆、多智能体编排与产品交付
03 /6+ hours saved weekly; product revenue ownership每周节省 6+ 小时，负责产品收入增长
04 /Six Sigma · Google · Machine Learning六西格玛 · Google · 机器学习
01 / Career focus求职方向
I’m focusing on three directions. Logistics, supply chains and process excellence remain the domain experience I bring to each.目前重点关注三个方向。物流、供应链与流程卓越，是我在这些方向上可迁移的业务积累。
Bring structure to ambiguous, cross-functional work. Align scope, dependencies, resources, cost and stakeholders—and stay accountable through delivery.为模糊、跨职能的工作建立结构，协调范围、依赖、资源、成本和干系人，并对最终交付负责。
Project / Program Management项目管理 · 项目群管理Connect product decisions with acquisition, activation, retention and revenue. Weigh delivery cost and profitability alongside growth, using data and user feedback to set priorities.把产品决策与获客、激活、留存及收入连接起来，同时考虑交付成本与盈利能力。用数据和用户反馈决定优先级，让增长落到实际业务回报上。
Growth Product Manager增长产品经理Work close to customers: understand the process, prototype an AI solution, integrate it into real systems, and support adoption. Bridge business context and engineering execution.贴近客户，理解流程，构建 AI 原型并接入真实系统，推动团队采用。在业务现场连接需求与工程执行。
FDE · Applied AI SolutionsFDE · 应用型 AI 解决方案02 / Selected work精选项目
Tools people can use. Processes teams can run. A selection of product and operational work.让人用得上的工具，让团队跑得通的流程。这里是我的部分产品与运营实践。
01 / Agent infrastructure智能体基础设施
Give the next session
a better starting point.让下一次对话，
从已有经验开始。
AI agents lose context between sessions. I built a local-first memory system that carries knowledge and corrections forward, with tools to measure whether the same mistakes recur.AI 智能体常常在会话之间丢失上下文。我构建了本地优先的记忆系统，保留知识与纠正，并提供测量工具，追踪同类错误是否再次发生。
My role: product direction, implementation, evaluation & releases.我的工作：产品方向、工程实现、评测与版本发布。
Explore the repository查看开源项目An agent can retrieve an old correction without changing its behavior. Memory needs both persistence and measurement.智能体能检索到过去的纠正，不代表行为真的改变。记忆既需要持久化，也需要度量。
A corrections ledger, session lifecycle, local memory layers and a dashboard. MCP, SDK and CLI make the system usable across development tools.纠正记录、会话生命周期、本地分层记忆与仪表盘。通过 MCP、SDK 和 CLI 接入不同开发工具。
The public repository includes reproducible evaluation artifacts and their limitations. The measured behavior-change uplift is still an open research question.公开仓库包含可复现的评测材料及其局限。记忆究竟能带来多大行为改善，仍是需要验证的研究问题。
From idea to a revenue-generating product.从想法，到产生收入的产品。
02 / Product × growth产品 × 增长
Build it. Launch it. Grow it.做出来，上线，再推动增长。
Built an AI service from scratch with Next.js, AWS and Claude API integration. My work connects product delivery with European acquisition and retention: turning user needs into priorities, and product usage into business value.用 Next.js、AWS 与 Claude API 集成，从零构建 AI 服务。我的工作连接产品交付与欧洲市场的获客、留存：把用户需求转成产品优先级，把产品使用转成业务价值。
Commercial product at Novada Labs · implementation is private.Novada Labs 商业产品 · 工程实现未公开。
My product & growth work我的产品与增长工作saved per week on team reporting团队每周节省的报表时间
03 / Operations × automation运营 × 自动化
Automation inside a real operation.在真实运营中，让自动化落地。
At Amazon, I built an AI-assisted reporting tool using Claude, VBA and Tampermonkey. Adopted by the team, it saved 6+ hours a week alongside my work on KPIs and process improvement.在亚马逊，我用 Claude、VBA 与 Tampermonkey 构建 AI 辅助报表工具。团队采用后每周节省 6 小时以上，与我的 KPI 管理和流程改进工作相互配合。
Explore the illustrative demo查看演示页面Reconstructed demonstration; not a live Amazon system.重建演示页面，并非亚马逊实时系统。
More things I’ve built更多我做过的东西
Agent instructions智能体指令
Turns goals into structured task briefs and repeatable SOPs. I built the compiler, routing logic and validation workflow to reduce ambiguity in complex agent tasks.把目标转成结构化任务说明与可复用 SOP。我构建了编译器、任务路由和验证流程，减少复杂智能体任务中的指令歧义。
Multi-agent orchestration多智能体编排
Coordinates workers, independent review and verification across multi-step tasks. Budget controls, checkpoints and escalation make long-running work easier to manage.在多步骤任务中协调执行者、独立审查与验证。通过预算控制、检查点和异常升级，让长时间运行的任务更可管理。
Voice product语音产品
A local-first macOS dictation app: on-device speech recognition, mixed Chinese and English, and system-wide text input. Built around a real daily workflow.本地优先的 macOS 听写应用：设备端语音识别、中英混合输入、全局文本输入。从真实的日常使用流程出发设计。
AI integrationsAI 系统集成
An integration layer that brings web data and tools into agent workflows. Covers the practical work of interfaces, tool behavior, testing and releases.把网络数据与工具接入智能体工作流的集成层，覆盖接口设计、工具行为、测试与发布等落地工作。
Supply-chain operations供应链交付管理平台
At Xiaomi, built company-wide KPI tracking and exception-management workflows: dashboards, delay and damage alerts, and escalation processes for cross-border fulfilment.在小米构建公司级 KPI 跟踪与异常管理流程：仪表盘、延误与破损预警、异常升级，服务跨境履约。
Early AI research · 2023早期 AI 研究 · 2023
Built document-question-answering experiments with LangChain and Pinecone during my Freiburg seminar work on the GPT ecosystem.在弗莱堡研究 GPT 生态期间，用 LangChain 与 Pinecone 构建文档问答实验，将 AI 研究转化为实际应用。
Earlier utility: 其他早期工具：Proxy4AgentProxy4Agent — a small MCP proxy integration. —— 轻量 MCP 代理集成。
03 / The connecting thread我的主线
Business fluency.
Technical curiosity.
End-to-end
ownership.理解业务。
探索技术。
对交付负责。
The useful part of my background is the combination. I understand the operational problem, translate it into a product, and stay with it through delivery.我的优势来自经历的组合：理解运营中的真实问题，把它转化为产品，并推动它走到交付。
Economics and finance taught me to question assumptions. Supply chains taught me that a process only works if people can actually run it. Building AI products brings those lessons together.经济与金融训练让我习惯检验假设。供应链让我明白，一套流程只有被人真正执行起来才有价值。构建 AI 产品，把这两种经验接在了一起。
Requirements, KPIs, root causes and the people behind the process.梳理需求、指标与根因，也理解流程背后的人。
AI workflows, agent memory, integrations and full-stack implementation.AI 工作流、智能体记忆、系统集成与全栈实现。
Cross-functional delivery, adoption, feedback and continuous improvement.跨团队交付、推动采用、收集反馈并持续改进。
04 / Experience完整经历
AI products, European operations, and a foundation in analytical thinking.AI 产品、欧洲运营经验，以及分析与研究的底子。
Own AI products and projects end to end—from requirements and prototypes to engineering, launch and iteration. Drive European product-revenue growth through acquisition and retention, aligning product, engineering and marketing. Coordinate a China-based product and engineering team and external stakeholders.端到端负责 AI 产品与项目，从需求、原型到工程、上线与迭代。围绕获客和留存推动欧洲产品收入增长，协调产品、工程与市场，并对接中国的产品技术团队及外部伙伴。
Mar 2026 — Present2026.03 — 至今Owned in-station and out-of-station process performance, KPI analysis and cost control. Worked across 16 carrier accounts and 100+ metrics; ran root-cause analysis, corrective actions, training and monthly audits. Built AI-assisted reporting adopted by the team, saving 6+ hours per week.负责站内与站外流程表现、KPI 分析和成本控制，覆盖 16 个承运商账户与 100+ 项指标。推进根因分析、纠正措施、培训及月度审计；团队采用 AI 辅助报表工具后，每周节省 6 小时以上。
Sep 2025 — Jan 20262025.09 — 2026.01Managed end-to-end fulfilment across Germany and the Benelux. DE/NL delivery scope included approximately €110M and 320,000 units per year at 99.5% on-time delivery. Built KPI and exception-management workflows; coordinated key accounts, 3PL partners, cost reporting, PSI and CPFR planning across 25+ product models.负责德国与比荷卢的端到端履约，其中德国和荷兰交付规模约为每年 1.1 亿欧元、32 万件，准时率 99.5%。构建 KPI 与异常管理流程，协调大客户、第三方物流、成本报告，以及 25+ 产品型号的 PSI 与 CPFR 计划。
Jun 2024 — Aug 20252024.06 — 2025.08Supported Western European certification, audit readiness and closed-loop finding management across ISO 9001, 14001 and 45001, with ESG and information-security compliance support. Coordinated product teams, laboratories and external stakeholders across languages.支持西欧认证、审计准备及问题闭环管理，涉及 ISO 9001、14001、45001，并协助 ESG 与信息安全合规；跨语言协调产品团队、实验室和外部伙伴。
Dec 2023 — Jun 20242023.12 — 2024.06Taught Advanced Macroeconomics I & II, Principles of Finance, and Economic Policy & Public Choice. Developed Python coursework, led classes and assessed exams. In parallel, used R and Z-tree for causal and behavioral-economics research.承担高级宏观经济学 I/II、金融学原理、经济政策与公共选择等课程教学，编写 Python 作业、带课并参与考试评阅；同时使用 R 与 Z-tree 开展因果与行为经济学研究。
Oct 2022 — Feb 20242022.10 — 2024.02Supported SAP procurement workflows, internal communication and internal-audit preparation in an automotive supply-chain environment.在汽车供应链环境中支持 SAP 采购流程、内部沟通及内部审计准备。
Jul 2021 — Oct 20212021.07 — 2021.10Ran an independent service practice: client communication, curriculum design, teaching, scheduling and delivery. Adapted learning plans to different students and managed the full customer relationship.独立经营教育培训与翻译服务，负责客户沟通、课程设计、教学、排期和交付；按不同学生的需要调整学习计划，管理完整客户关系。
Mar 2019 — Sep 20202019.03 — 2020.09Supported overseas B2B business development and export fulfilment: sea, air and road freight coordination, 3PL partners, shipping documentation and cross-border customs clearance.支持海外 B2B 客户开发与出口履约，协调海陆空运输、第三方物流、出货单证及跨境清关。
Oct 2018 — Mar 20192018.10 — 2019.03Supported retail logistics, sales and daily operations in a high-throughput US retail environment.在美国高吞吐量零售环境中参与物流、销售与日常运营。
Jun 2018 — Sep 20182018.06 — 2018.09Education & research教育与研究
2020 — 2023
M.Sc. Economics & Finance经济与金融学硕士
Computational economics, machine learning, econometrics and time-series analysis. Outstanding student representative and commencement speaker.计算经济学、机器学习、计量经济学与时间序列分析。优秀学生代表、毕业典礼致辞代表。
GPT ecosystem paperGPT 生态论文Master’s thesis硕士论文2014 — 2018
B.A. Business English · International Trade & Business商务英语学士 · 国际贸易与商务
International trade and finance, economics, management and languages. Vice president of the 13th Student Union; outstanding graduate representative.国际贸易与金融、经济学、管理及语言。第十三届学生会副主席、优秀毕业生代表。
Python · TypeScript · SQL · R · Power BI · PostgreSQL · MCP · RAG · Multi-agent workflows · LLM integration.Python · TypeScript · SQL · R · Power BI · PostgreSQL · MCP · RAG · 多智能体工作流 · 大模型集成。
Lean / 5S · Six Sigma Green Belt · Root-cause analysis · KPI design · CPFR · Supply-chain planning · Cross-functional project management.精益 / 5S · 六西格玛绿带 · 根因分析 · KPI 设计 · CPFR · 供应链计划 · 跨部门项目管理。
05 / Credentials证书与资质
Selected completed programs in process improvement, analytics, AI and business analysis. Each links to its certificate record.流程改进、数据分析、AI 与业务分析领域的课程与专业证书，均附证书核验入口。
Process improvement流程改进
Kennesaw State University · Coursera Specialization肯尼索州立大学 · Coursera 专项课程结业证书
Completed 19 January 2026 · 4 courses完成于 2026 年 1 月 19 日 · 4 门课程
Define → Measure → Analyze → Improve → Control. A structured approach to variation, quality, root causes and operational cost.定义 → 测量 → 分析 → 改进 → 控制。用结构化方法处理波动、质量、根因与运营成本。
Professional Certificate · 8 courses专业证书 · 8 门课程
Data preparation, SQL, visualization and R.数据准备、SQL、可视化与 R。
View certificate核验证书Professional Certificate · 6 courses专业证书 · 6 门课程
Python, Git, debugging and automation at scale.Python、Git、调试与规模化自动化。
View certificate核验证书Specialization · 3 courses专项课程 · 3 门课程
Supervised learning, neural networks and recommender systems.监督学习、神经网络与推荐系统。
View certificate核验证书Specialization · 5 courses专项课程 · 5 门课程
The PwC Approach: from business data to decisions and communication.PwC 方法：从业务数据到决策与沟通。
View certificate核验证书Specialization · 4 courses备考专项课程 · 4 门课程
Requirements, stakeholder engagement and solution evaluation. Course completion; not the IIBA CBAP designation.需求管理、干系人协作与解决方案评估。课程结业证书，非 IIBA CBAP 资格认证。
View certificate核验证书Specialization · 4 courses专项课程 · 4 门课程
ESG risks, investment perspectives, climate and corporate responsibility.ESG 风险、投资视角、气候与企业责任。
View certificate核验证书Chinese · native / English · C2 / German · C1中文 · 母语 / 英语 · C2 / 德语 · C1
EU Blue Card · Based in Düsseldorf, Germany欧盟蓝卡 · 常驻德国杜塞尔多夫
