Contact

Profile

Engineer focused on AI applications and technical innovation, with a solid background in machine learning and deep learning. I specialize in turning cutting-edge technology into deployable, scalable systems.

Cathay Financial Holdings (OCR & Document Intelligence)

At Cathay, I led the build of a medical receipt OCR system: data preprocessing, Traditional Chinese PaddleOCR training and optimization, and table/text parsing. The system improved recognition accuracy and throughput and established a scalable intelligent document processing architecture.

Recent Focus (LLM, Agent, RAG, AIGC)

I now focus on LLM applications and AI IDE–driven development, with an emphasis on speed and security. Highlights include:

Introducing LLM as a Judge on a feedback platform to turn qualitative feedback into quantitative scores
Building an Agentic RAG architecture for the M365 knowledge assistant to support real-time customer support and knowledge retrieval
Building a conference news and trend summarization tool
Implementing a digital human video application to extend AIGC use cases

Current Research (Agent, MCP, Enterprise Architecture)

I am deepening work on AI Agent design and Model Context Protocol (MCP)—multi-agent coordination, tool orchestration, and safety boundaries—and aim to:

Refactor development workflows with AI IDE
Balance security and system stability with fast iteration
Establish sustainable enterprise-grade AI architecture

Value Proposition

I believe the value of AI lies not only in model capability but in how we integrate technology, process, and safety with an engineering mindset to deliver efficient, scalable intelligent applications.

Open to discussions on AI platform, Agent, RAG, or Realtime AI—consulting, development, or full-time roles.

LinkedIn GitHub

Looking for an AI platform or Agentic AI partner? Let's take GenAI from PoC to production.

Contact on LinkedIn

簡介

熱衷於 AI 應用與技術創新的工程師，具備扎實的機器學習與深度學習背景，專注於將前沿技術快速轉化為可落地、可擴展的實務系統。

國泰金控時期（OCR 與文件智能）

加入國泰金控後，主導醫療收據 OCR 系統建置，負責資料前處理流程設計、繁體中文 PaddleOCR 模型訓練優化與表格文字解析演算法開發，成功提升醫療單據辨識準確率與處理效率，建立可擴展的智能文件處理架構。

近年重心（LLM、Agent、RAG、AIGC）

近年專注於大型語言模型（LLM）應用與 AI IDE 全新開發模式，強調高速開發與安全並行。實務成果包含：

於回饋平台導入 LLM as a Judge，將質化回饋轉化為量化評分
於 M365 知識助手建構 Agentic RAG 架構，支援業務即時客服與知識檢索
打造研討會新聞趨勢整理工具
實踐數字人影片製作應用，拓展 AIGC 落地場景

當前研究方向（Agent、MCP、企業架構）

目前深入研究 AI Agent 架構與 Model Context Protocol（MCP），探索多代理協作、工具調度與安全邊界設計，致力於：

以 AI IDE 重構開發流程
在高速迭代中兼顧資安與系統穩定性
建立可持續演進的企業級 AI 架構

價值主張

我相信 AI 的價值不僅在模型能力，而在於如何以工程思維整合技術、流程與安全機制，真正實現高效率且可規模化的智慧應用。

聯絡方式

歡迎洽談 AI 平台、Agent、RAG、Realtime AI 相關的顧問或開發合作、以及求職機會。