Hello I'm Charles Tsao.
Senior Data Science Analyst
Based In Taipei, Taiwan.
I specialize in machine learning, LLM-driven multi-agent systems, OCR automation, and GraphRAG retrieval engineering. I build end-to-end AI solutions — spanning modeling, explainability, deployment, and enterprise-scale integration — with a primary focus on insurance claims analytics, fraud detection, and workflow automation.
My Skills
Python
Machine Learning (LightGBM, XGBoost)
SHAP Explainability
LLM / OpenAI
Multi-Agent Workflow Design
GraphRAG / Vector Search
OCR / Computer Vision
FastAPI
SQL
PySpark
CI/CD for ML Deployment
Git / GitHub Actions
Data Pipeline Design
Technical Project Management
Cross-Functional Collaboration
Python
Machine Learning (LightGBM, XGBoost)
SHAP Explainability
LLM / OpenAI
Multi-Agent Workflow Design
GraphRAG / Vector Search
OCR / Computer Vision
FastAPI
SQL
PySpark
CI/CD for ML Deployment
Git / GitHub Actions
Data Pipeline Design
Technical Project Management
Cross-Functional Collaboration
My Experience
Senior Data Science Analyst
Cathay Life Insurance
Nov 2019 - PresentLead ML/AI initiatives across claims analytics, fraud detection, and LLM automation. Deliver full-cycle AI solutions from modeling to deployment.
- Developed fraud detection models and risk scoring systems to identify suspicious patterns.
- Designed SHAP-based explainability for regulatory compliance.
- Led OCR POCs with startups to improve medical invoice extraction.
- Built LLM multi-agent workflows integrating business rules and vector search.
- Implemented GraphRAG to enhance internal retrieval precision.
- Managed enterprise LLM Employee Assistant roadmap and rollout.
- Delivered complete ML lifecycle: data → modeling → CI/CD → monitoring.
About Me
I am a data science professional dedicated to building intelligent systems that solve real operational challenges. My work focuses on integrating machine learning, large language models, graph-based reasoning, and automation to improve efficiency and strengthen risk control in highly regulated industries.
I have hands-on experience developing fraud detection models, OCR pipelines, multi-agent decision systems, and GraphRAG retrieval architectures — and successfully deploying them into production environments. I enjoy partnering with IT teams, business operations, actuaries, and medical experts to translate complex domain logic into scalable AI workflows.
With a hybrid background across data science, LLM engineering, and product management, I aim to create AI solutions that are technically robust, explainable, and truly impactful for end users.
My Projects
02
Medical OCR Pipeline (Startup Collaboration)
A POC with external startups to build OCR solutions for medical invoices and
certificates. Combined deep-learning OCR with structured rule extraction to enhance
downstream claims processing.
Impact: Significantly increased structured field accuracy and reduced manual data
entry workload.
Let's talk for
Something special
I’m open to opportunities in ML engineering, LLM system development,
RAG/GraphRAG, and enterprise AI applications.
Feel free to reach out — let’s build something impactful together.
Taipei, Taiwan