AI Agent Developer
刘炳中
8 years building AI agents, backend systems, and ML infrastructure. Python / Go. From multi-modal computer-use agents to production RAG systems — passionate about embedding AI into real-world workflows at scale.
Technical Expertise
AI Agent Systems
- Agent harness design patterns, Plan→Act→Eval→Memory loop
- LangGraph multi-step reasoning, MCP tool-calling protocol
- Computer Use multi-modal agents (Qwen3VL)
- Skill plugin system (register, hot-load, visual orchestration)
Backend Systems
- Python (FastAPI, Flask) & Go (gRPC, Gin, Fiber)
- Distributed microservices & event-driven architecture
- Celery, RabbitMQ, Kafka, Ray for async & batch
- PostgreSQL, MySQL, MongoDB, Redis, Elasticsearch
ML & Infrastructure
- vLLM inference deployment, SFT / RL fine-tuning
- RAG hybrid retrieval (pgvector + ES + BGE Re-rank)
- Docker, K8s, CI/CD, Cloudflare Workers
- Prometheus, Grafana observability
Featured Projects
Go-ComputerUseAgent
High-performance automated computer-use agent combining Go's performance with intelligent autonomous systems.
Go · AI Agent Architecture · Multi-modalstar-manager
Automated GitHub Stars classification — fetches, classifies, and organizes starred repositories into categorized GitHub Lists.
Python · GitHub Actions · Automation13340146002 / bzfq21 (微信) / bzfq21@gmail.com / mail.bzfq21.qzz.io