5+ years building backend systems across .NET and Python (FastAPI). Currently leading the technical transformation of AI products from .NET legacy to Python / Cloud-Native architectures, with a focus on LLM integration and AI-driven development workflows.

• Working on: Text2SQL engine for cross-database natural language querying (MySQL / PostgreSQL / MSSQL)
• Exploring: Advanced RAG patterns, multi-agent systems, MCP protocol
• AI Stack: Production LLM applications on Azure OpenAI; running POCs on AWS Bedrock & GCP Vertex AI
• Dev Workflow: All-in on Cursor & Claude Code for AI-assisted development

Languages & Frameworks

AI & LLM

Cloud & DevOps

Databases & Data

Production-ready v1.0 · 17-agent multi-collaborator architecture for portfolio rebalancing decisions targeting 15% annual returns. Built with FastAPI, React, PostgreSQL, Redis, and ML models (LSTM / XGBoost / LightGBM). Integrates Taiwan-specific regulatory costs and multi-source data resilience.

In-house Text2SQL engine supporting MySQL, PostgreSQL, and MSSQL. Multi-LLM switching strategy with Azure OpenAI (primary) and AWS Bedrock / GCP Vertex AI (POC) to balance accuracy and token cost. Schema-as-Vector retrieval reduces token usage by 92%.

Reverse-engineered 2 outsourced C# projects without documentation. Combined Swagger API probing with Cursor-assisted analysis to reconstruct architecture, generate functional specs, and migrate legacy logic precisely to Python 2.0.

Engineered backend APIs and payment systems for 200+ units nationwide across Taiwan. Integrated Credit Cards, iCash, LinePay, and Open Point digital wallets. Increased deployment efficiency by 30% and reduced financial accounting workload by 50%.

• Email: [email protected]
• Location: Taipei, Taiwan 🇹🇼
• Status: 🟢 Open to Senior Backend / AI Engineering opportunities

"The best engineers don't just adapt to change—they lead it."