About me關於我

I specialize in translating AI into practical SaaS products. Covering the complete lifecycle from model to launch, I ensure every technical investment yields a positive return.我專注於將 AI 轉化為實際的商業 SaaS 產品，包辦從模型建置到系統上線的完整週期，確保每一項技術投入都能創造超越成本的實質收益。

I'm a data scientist and AI solutions developer with over eight years of experience building systems that put machine learning into everyday business decisions. My PhD, from the University of Manchester, focused on natural language processing and marketing analytics.

Most recently I was Data Science & AI Manager at FTI Consulting in Hong Kong, where our team served as an embedded data science function for client organisations — accountable not for the analysis itself, but for the decisions it was meant to improve.

Across more than twenty projects in financial services, e-commerce, and technology, two illustrate the range. For a global tier-one e-commerce platform, I built a crisis forecasting model that gives communications and business leaders three to six hours of warning before an issue surfaces publicly. For a Hong Kong financial institution, I designed an anti-money-laundering transaction monitoring system that won a Regulation Asia Award for Excellence in 2023.

我是一名資料科學家與 AI 解決方案開發者，擁有超過八年經驗，專注於打造能將機器學習融入日常商業決策的系統。我在曼徹斯特大學取得博士學位，研究領域聚焦於自然語言處理與行銷分析。

最近，我在香港的 FTI Consulting 擔任資料科學與 AI 經理。我所帶領的團隊以客戶組織內嵌資料科學團隊的形式合作；我們負責的不只是分析本身，更要對分析所要改善的決策負責。

在金融服務、電子商務與科技產業累積的二十多個專案中，有兩個案例展現了工作的廣度。針對一家全球一線電子商務平台，我建立了一套危機預測模型，讓企業傳訊與業務領導者能在事件公開浮現前三至六小時收到預警。針對一家香港金融機構，我設計了一套反洗錢交易監控系統，並於 2023 年獲得 Regulation Asia Awards for Excellence 獎項。

Experience工作經歷

2023 – 2025

Director, Data Science & AI Solution

FTI Consulting · Hong KongFTI Consulting · 香港

Led 20+ enterprise AI and data analytics initiatives across financial services, legal, HR, and corporate communications — translating ambiguous business needs into prioritised roadmaps, scalable architectures, and production workflows.主導 20 多項企業 AI 與數據分析專案，橫跨金融服務、法務、人力資源與企業傳訊等領域——將模糊的客戶需求，轉化為具優先順序的發展藍圖、可擴展的架構，以及正式上線的工作流程。
Designed and deployed an LLM-powered talent recommender using GPT-4, Llama 3, and RAG, serving 8,000+ corporate users at production-grade latency.設計並部署一套採用 GPT-4、Llama 3 與 RAG 技術的 LLM 人才推薦系統，服務超過 8,000 名企業使用者，達到正式環境等級的延遲表現。
Led end-to-end re-engineering of a legacy transaction monitoring platform for a Hong Kong brokerage, improving stability and scalability while cutting monthly operating cost from USD 200K to USD 30K.主導一家香港券商舊有交易監控平台的全面重構，提升系統穩定性與可擴展性，同時將每月營運成本從 20 萬美元降至 3 萬美元。
Architected an AI-powered financial crime compliance system for a global technology company — winner of Best AML/CFT Data, Analytics & Visualisation Solution at the 2023 Regulation Asia Awards.為一家全球科技公司架構一套 AI 驅動的金融犯罪合規系統——榮獲 2023 年 Regulation Asia Awards「最佳反洗錢/打擊資恐數據、分析與視覺化解決方案」獎項。
Built an India-based support team for monitoring, incident detection, escalation, and service readiness, strengthening observability and continuity across regions.建立一支駐印度的支援團隊，負責監控、事故偵測、升級處理與服務就緒性，強化跨地區的可觀測性與營運持續性。
Designed agentic AI workflows with Ollama, LangChain, FastAPI, and Open WebUI for legal document analysis and contract generation, bridging the platform team and the corporate finance practice.運用 Ollama、LangChain、FastAPI 與 Open WebUI，設計代理式 AI（agentic AI）工作流程，用於法律文件分析與合約生成，銜接平台團隊與企業財務顧問業務。

2019 – 2023

Senior Consultant, Data Analytics

FTI Consulting · London, UK / Hong KongFTI Consulting · 英國倫敦 / 香港

Scoped and delivered a crisis analytics system combining foundation models with time-series anomaly detection, flagging critical events 3–6 hours earlier than a 150-person manual monitoring team at a global ecommerce company.界定範疇並交付一套結合基礎模型與時間序列異常偵測的危機分析系統，為一家全球電商公司，比一支 150 人的人工監控團隊提早 3 至 6 小時標示出關鍵事件。
Created a no-code analytics platform that standardised repetitive workflows and let non-technical teams run analyses independently, reducing project delivery time by 90%.打造一套無程式碼（no-code）分析平台，將重複性工作流程標準化，讓非技術團隊也能獨立執行分析，使專案交付時間縮短 90%。
Designed and operationalised an ontology-based KYC automation platform for a Canadian bank, reducing manual review by 60% while strengthening governance and data quality controls.為一家加拿大銀行設計並落地一套以本體論（ontology）為基礎的 KYC 自動化平台，將人工審查工作減少 60%，同時強化治理與資料品質控管。

2018 – 2019

NLP Data Scientist

Hello Soda · Manchester, UK — fintech startupHello Soda · 英國曼徹斯特——金融科技新創公司

Designed multi-task deep learning classifiers for customer-interest classification, combining text and behavioural data to improve prediction accuracy by 30%.設計多任務深度學習分類器，用於客戶興趣分類，結合文字與行為資料，將預測準確度提升 30%。
Deployed XGBoost sentiment models into low-latency production, using A/B testing and continuous evaluation to monitor performance and model drift.將 XGBoost 情緒分析模型部署至低延遲的正式環境，透過 A/B 測試與持續評估，監控模型表現與飄移狀況。
Optimised ML workflows from experimentation through deployment on AWS, focused on reducing inference latency for real-time customer scoring.優化從實驗到部署的機器學習工作流程（於 AWS 上執行），著重於降低即時客戶評分的推論延遲。

2014 – 2016

Research & Teaching Assistant

The University of Manchester · Manchester, UK曼徹斯特大學 · 英國曼徹斯特

Delivered seminars on big data, statistics, and data mining to undergraduate and postgraduate students.為大學部與研究所學生講授大數據、統計學與資料探勘課程。
Authored user manuals and case studies for SAS and IBM SPSS Modeler.撰寫 SAS 與 IBM SPSS Modeler 的使用手冊與案例研究。
Presented research and applied work at 10+ academic and industry conferences.於 10 多場學術與產業研討會發表研究與應用成果。

Core expertise核心專長

AI & digital transformationAI 與數位轉型: Enterprise AI strategy, business process re-engineering, workflow automation, change enablement.企業 AI 策略、業務流程再造、工作流程自動化、變革推動。
Solution architecture & delivery解決方案架構與交付: End-to-end AI product lifecycle, data processing, API integration, workflow orchestration, production deployment, monitoring, governance.AI 產品全生命週期管理、資料處理、API 整合、工作流程協同、正式環境部署、監控、治理。
AI / MLAI / ML: Generative AI, RAG, NLP, predictive analytics, anomaly detection, time-series modelling, explainable AI, human-in-the-loop systems, TensorFlow, PyTorch, scikit-learn.生成式 AI、RAG、自然語言處理（NLP）、預測分析、異常偵測、時間序列建模、可解釋 AI（explainable AI）、人機協作系統（human-in-the-loop）、TensorFlow、PyTorch、scikit-learn。
Programming程式設計: Python, JavaScript, SQL.Python、JavaScript、SQL。
BI & analytics商業智慧與分析: Tableau, Power BI, ReactJS, Streamlit.Tableau、Power BI、ReactJS、Streamlit。

Education教育背景

2013 – 2018

PhD in Data Science, Business and Management數據科學暨商業管理博士

The University of Manchester, UK曼徹斯特大學，英國

2011 – 2012

MSc in Marketing行銷碩士

The University of Manchester, UK曼徹斯特大學，英國

2006 – 2010

BA in Radio and Television廣播電視學士

National Chengchi University, Taiwan國立政治大學，台灣

Talks公開演講

2020Apply semantic analysis and deep learning for corporate crisis detectionPyData Taipei
2018Shift from multi-class classification model to multi-task learning networkPyData MCR
2018Unbox the magic of natural language processingHer + Data Manchester

Certifications專業認證

AWS Certified Machine Learning — Specialty
AWS Certified Developer — Associate