yaogdu - Overview
👋 Hi, I'm Toby Du
Better late than never.
🚀 About Me
I'm a backend architect and engineering leader with 10+ years of experience building distributed systems, observability platforms, and enterprise-grade AI infrastructure.
- 🔭 Currently focused on AI-powered tools, AgentOps, RAG workflows, and the MCP ecosystem
- 🏗️ Experienced in designing and scaling distributed systems supporting 10M+ DAU
- ⚡ Passionate about high concurrency, high availability, reliability engineering, and developer efficiency
- 👥 7+ years of experience in team leadership, technical governance, and cross-functional delivery
- 💡 Interested in knowledge management, productivity tools, fintech, and developer platforms
- 📍 Based in Beijing, China
🧠 What I Do
- Build high-performance backend systems with strong stability and observability
- Design cloud-native architectures for large-scale business platforms
- Develop enterprise AI capabilities with LangGraph, AutoGen, MCP, RAG, and AgentOps
- Bridge technical architecture and product goals to deliver systems that create real business value
🛠️ Tech Stack
Languages
Backend & Infra
Cloud & Observability
AI Engineering
🏆 Selected Highlights
- Scalable Architecture Expert — 10+ years in backend engineering, with hands-on experience designing distributed systems for 10M+ DAU
- Reliability Driven — built and scaled platforms with 99.9%+ availability, strong observability, and minute-level alerting
- Polyglot Engineer — advanced proficiency in Java (Spring Boot), Go, and Python
- Technical Leadership — 7+ years leading teams, improving R&D workflows, and building engineering governance
- AI Infrastructure Builder — experienced in enterprise-grade Agents, RAG workflows, MCP integrations, and AgentOps platforms
- Product-Minded Architect — able to turn complex technical requirements into practical systems that drive business outcomes
💼 Experience Snapshot
- Orange Digital Technology — Led architecture design and built the company-wide observability stack with Prometheus, SLS, Grafana, and N9E, achieving 99.9% uptime and minute-level monitoring
- Huanxin Network — Built the business middle platform from scratch, supporting 6 business lines, millions of DAU, and 99.99% system stability
- JD Technology — Supported large-scale marketing and coupon systems during major promotional campaigns with millions of QPS
- VIPKID — Led booking and class-management architecture, supporting 100,000 peak QPS and 10+ collaborating teams
- Duolabao — Designed core payment systems, OAuth2 open platform, data synchronization middleware, and big data capabilities
📦 Featured Projects
SkillHub: Enterprise AI Asset Distribution & AgentOps Infrastructure
A unified distribution system for Prompts, Agent scripts, and MCP Servers, built around the npx interaction paradigm with enterprise-grade enhancements.
- Standardized package format for Prompts / Agents / MCPs
- Automated runtime isolation with Python venv and Node modules
- Shadow path mapping via symlink orchestration
- Deep integration with Anthropic MCP
- Offline-ready versioned snapshotting and metadata indexing
- Static auditing pipeline for high-risk command filtering
Enterprise Stability & Observability Revitalization
A company-wide observability initiative designed to shift engineering from reactive firefighting to proactive system management.
- Built a unified observability platform across Java, Go, and Python
- Developed an AI-driven Alert Gateway with alert convergence and automated ACK workflows
- Implemented graceful startup/shutdown protocols to reduce deployment failures
- Drove adoption of standardized monitoring and on-call governance
- Reached 85% monitoring coverage and minute-level anomaly detection
📈 GitHub Activity
💭 Random Dev Quote
📫 Let's Connect
- Email: yaogdu@gmail.com
- GitHub: github.com/yaogdu
- Open to collaborating on backend architecture, AI infrastructure, observability, and developer tooling