05/30/2026 updated
100 % available
AI Forward Deployed Engineer/Multi-Agent AI Systems, MCP Tools, LLM Automation & Full-Stack Delivery
ShenZhen, China
Worldwide
MSc Epidemiology and Biostatistics, School of Public Health, Sun Yat-sen UniversityAbout me
I am an AI FDE who turns business problems into working AI products. I build multi-agent systems, MCP tools, workflow automation and full-stack platforms, with a strong focus on real deployment, user value and fast iteration. Founder and builder of PigBun-AI, an AI platform focused on AI co-founder.
AI AgentsLangChainPython (Programming Language)PostgreSQLMessaging Application Programming InterfaceTypeScriptLarge Language ModelsMulti-Agent SystemsPrompt EngineeringFastAPIDeployment AutomationPlaywrightReactive ProgrammingFront End Software DevelopmentArtificial Intelligence Markup Language (AIML)
AI Forward Deployed Engineer specializing in real-world AI implementation, multi-agent systems, LLM automation, MCP tools and full-stack AI product delivery.
I help businesses turn manual, complex or unclear workflows into production-ready AI systems. My strength is not only building prototypes, but designing, deploying and iterating AI solutions that work in real business environments.
Core expertise:
- AI agents and multi-agent orchestration: LangGraph, LangChain, agent workflows, task decomposition, human review loops, agent handoff and status protocols
- LLM application development: dify, langgraph, prompt engineering, harness, tool calling, structured outputs, RAG and AI workflow design
- MCP development: Model Context Protocol servers, AI tool integrations, browser automation tools, knowledge graph memory and agent-accessible APIs
- Full-stack development: React, Next.js, TypeScript, Vite, FastAPI, Node.js, Python, PostgreSQL, Redis, MinIO, Neo4j, RabbitMQ, Docker and Docker Compose
- Browser and workflow automation: Playwright, OCR/PDF extraction, ERP automation, social media automation, data collection and operational dashboards
- AI media and vision systems: image analysis, video analysis, content classification, vector search, pgvector, Milvus, ComfyUI, GPU workflow deployment
- Product and delivery: requirements discovery, system architecture, MVP delivery, API design, frontend dashboards, backend services, deployment, testing and iteration
Selected experience:
- Built PigBun-AI / pigbunai.com, an AI platform with website, plugin registry, authentication gateway, deployment services, AI partner concepts and agent-team orchestration.
- Developed MCP tools for RedNote/Xiaohongshu research, publishing, analytics, multi-session automation and creator operations.
- Built knowledge graph MCP infrastructure using Neo4j to turn AI agent problem-solving experience into reusable decision paths.
- Delivered AI automation systems for ERP form filling, PDF/OCR extraction, creator content screening, AI VLOG editing, image collection and social media workflows.
- Worked on AI fashion/search platforms using LangGraph pipelines, vector search, image understanding, A100/ComfyUI infrastructure and multi-service backend systems.
- Built Docker-based ComfyUI cluster management with GPU isolation, queue-based task routing, Redis/MinIO integration and health checks.
- Applied agent-assisted research workflows in WorldQuant Brain and received recognition in the International Quant Championship 2026, Stage 1, Top 20% of Teams Globally.
Industries and use cases:
AI SaaS, creator economy, social media automation, fashion AI, e-commerce operations, ERP automation, content intelligence, media generation, industrial digital twin systems and quantitative research.
I am a practical AI FDE: I work close to the business problem, identify where AI can create real leverage, build the technical system, and make it usable in production.
Experience
Professional experience5 - 10 years
Freelancer since2023
Languages
ChineseNative speakerEnglishFluent
Project history
Tasks and responsibilities during the project:
Founded and built PigBun-AI, an AI partner product focused on agentic workflows, MCP tools, AI automation and practical AI implementation. Owned the full product and engineering lifecycle, including product strategy, system architecture, frontend, backend, agent orchestration, deployment and operations.
Responsibilities included:
- Designed and developed the PigBun-AI platform and website at pigbunai.com
- Built the full-stack product architecture, including frontend, backend services, authentication, plugin registry and deployment workflows
- Designed AI partner concepts and agent-team orchestration workflows for real business execution
- Built and integrated MCP tools, AI automation capabilities and workflow-oriented agent systems
- Developed product pages, dashboards, deployment services, operational workflows and internal protocols
- Managed infrastructure, Docker-based deployment, CI/CD, release operations and production iteration
- Acted as founder, product owner, architect, full-stack engineer, AI workflow designer and deployment lead
Tasks and responsibilities during the project:
Founded and built a fashion intelligence / fashion report platform from concept to deployment. Owned the full technical delivery, including product architecture, frontend and backend development, AI workflow design, deployment, data algorithms and the annotation system.
Responsibilities included:
- Designed and developed the full-stack platform for AI-driven fashion report generation and analysis
- Built backend services, frontend interfaces, data workflows and deployment infrastructure
- Developed data processing and algorithm logic for fashion analysis, tagging and report generation
- Designed and implemented an internal annotation system to support structured fashion data labeling
- Integrated AI/LLM workflows for fashion insight extraction, classification and content generation
- Managed database design, APIs, deployment, testing and iteration
- Led the project as founder, combining product direction, engineering execution and AI implementation
Website:
ai-moda.ai
Tasks and responsibilities during the project:
Led Aura Future Technology as CEO and CTO, building an AI-powered interactive story game platform that transforms novels into playable interactive stories. The product vision was to make interactive game creation as simple as editing a video.
Responsibilities included:
- Defined product strategy, technical architecture and AI-native game creation workflow
- Built a system for converting novel content into branching interactive story experiences
- Designed agentic story generation pipelines using LangGraph and PromptFlow
- Used Neo4j to model story structure, characters, relationships, scenes and narrative paths
- Integrated ComfyUI-based visual generation workflows for story assets and game content
- Trained image LoRA models for visual style consistency and character/content generation
- Led both product and engineering execution as CEO and CTO
- Represented the company as an exhibiting guest at the first Shanghai Artificial Intelligence Conference
- Selected among the global top 30 AI application companies for roadshow presentation
- Won champion of the 5th Shanghai Yiyou Youth Incubator program