AI Application & Backend Engineer based in Taiwan — turning AI capabilities into production-ready products.
I design and ship backend systems powered by real AI — from LLM-integrated APIs and voice synthesis pipelines to autonomous agent workflows. My stack spans C#, Python, and edge infrastructure on Cloudflare. I don't just write code. I own the full cycle from architecture to deployment.
I'm an AI Application Engineer and Backend Systems Builder passionate about creating real-world impact through technology.
My journey started in traditional backend engineering with C# and ASP.NET, building scalable microservices and distributed systems. Over the years, I've transitioned into the exciting world of AI, focusing on practical applications of LLMs, AI agents, and autonomous systems that solve real business problems.
Today, I specialize in bridging the gap between AI and production systems—designing architectures that are not just technically sound but also deployment-ready and maintainable at scale.
AI & ML
LLM Integration
AI Agents
Voice AI
Data Processing
Backend
C#/ASP.NET
Python
Microservices
Event-Driven Architecture
Cloud & DevOps
Kubernetes
Cloudflare
Docker
CI/CD Pipelines
🎯
Execution-Focused
I ship, not just ideate
🏗️
System Builder
Design for scale and reliability
🧠
AI Native
Modern tech first
📚
Knowledge Sharing
Education through content
Featured Projects
Showcasing my recent work in AI, backend systems, and cloud infrastructure
AI
GPT-SoVITS Voice Cloning System
A production-grade voice synthesis pipeline built on GPT-SoVITS, supporting real-time cloning from short audio references with multi-language output and WebSocket-based streaming delivery.
Problem Solved
Most open-source TTS tools require hours of labeled audio and complex fine-tuning before producing usable output. Reduced voice reference requirement to under 30 seconds of audio, achieving sub-500ms initial response latency over WebSocket.
A multi-agent trading system that combines LLM-based reasoning with real-time market data to automate signal generation, risk assessment, and portfolio-level execution decisions.
Problem Solved
Manual trading analysis is slow, emotionally biased, and unable to process multiple data streams in parallel. Rule-based automation tools lack the contextual reasoning required at the decision layer. Full decision traceability enables post-analysis and strategy iteration without black-box guesswork.
A production microservices platform built with ASP.NET Core, featuring API gateway routing, event-driven messaging, and distributed tracing — designed for independent scaling and fast, low-risk deployments.
Problem Solved
A monolithic system was creating deployment bottlenecks, uneven load scaling, and tight coupling between business domains. Enabled per-service independent deployment and service-level failure isolation to prevent cascading outages across the platform.
In the age of AI, being discoverable by LLMs and search engines is crucial for thought leadership. This website demonstrates best practices in AI-optimized content architecture.
Semantic Structure
Well-organized HTML with clear heading hierarchy, semantic tags, and structured data for LLM comprehension.
Content Strategy
Original, deep-dive technical content that showcases expertise and attracts AI citations from research papers.
Technical SEO
Fast-loading, crawlable content optimized for both traditional search engines and AI language models.
Thought Leadership
Original insights and unique perspectives that stand out and get referenced by AI systems.
This Website: Our Case Study
This portfolio demonstrates AI visibility optimization through semantic HTML, clear information architecture, and high-quality technical content designed for both human readers and AI language models to understand and cite.
Result: Optimized for discovery by research AI, ChatGPT, Claude, and other LLMs through structured, original technical content.
Client Case Studies (Coming Soon)
Detailed case studies showing how AI visibility optimization has helped clients increase organic discovery, improve thought leadership positioning, and attract citations from AI systems.
Get In Touch
I'm always interested in discussing AI applications, backend systems, and new opportunities. Feel free to reach out!