Ming Fang
Technology Consultant & AI Solutions Architect
Building intelligent systems that process, understand, and transform complex data into actionable outcomes.
Featured Work
Multilingual Audio Processing Platform
A complete audio intelligence pipeline for language education. From multi-speaker recordings to isolated segments in under 3 minutes.
Key Results
- 100% segmentation accuracy (33/33 signal points)
- 40x faster processing (2+ hours → <3 minutes)
- <$0.50 per file (fully automated)
Tech Stack
Technical Capabilities
AI & Machine Learning
- Speaker Diarization
- Transcription
- LLM Integration
- Signal Processing
Full-Stack Development
- Next.js 15
- React 19
- Python FastAPI
- TypeScript
Audio/Video Processing
- Web Audio API
- ffmpeg
- wavesurfer.js
- Real-time Processing
Infrastructure
- AWS Lambda
- Vercel
- PostgreSQL
- Cloudflare
Latest Writings
The Complete Guide to Agent Configuration: How I Built a Self-Optimizing Workspace for My AI Agent
Most people treat AI agents like chatbots because their workspace files are broken. Here's how to fix that — with real before/after examples from a day of restructuring my own agent's configuration.
March 17, 2026
We Researched Every LLM Eval Tool So You Don't Have To
We compared LangFuse, Braintrust, and Datadog for LLM eval. Here's what we chose and a 3-stage guide to build your own eval pipeline in one week.
May 17, 2026
My OpenClaw Multi-Agent Workflow: From One AI to a Full Team
How I use OpenClaw to run parallel AI agents, 90+ skills, 6 real workflows, and copy-paste prompts you can use today.
March 19, 2026
View All Posts→Ventures
Enterprise Experience
Let's build something together.
Available for consulting, technical advisory, and select projects.