AI Algorithm Engineer

• 🏢 目前就职于 中云微迅科技有限公司，担任AI算法工程师

  Currently working at Zhongyun Weixun Technology Co., Ltd. as AI Algorithm Engineer

• 🎯 专注于 AI Agent平台、大模型工具链、推荐系统 和 大规模微服务架构

  Focused on AI Agent Platform, LLM Toolchain, Recommendation Systems and Large-scale Microservices Architecture

• 💡 擅长构建从0到1的AI产品，主导过省文旅平台AI Agent产品技术开发

  Expert in building AI products from scratch, led the development of provincial cultural tourism AI Agent platform

• 📊 参与过亿级流量推荐系统工程开发，经历APP日活从0到2300万的增长

  Experienced in billion-scale recommendation system engineering, grew app DAU from 0 to 23M

• 📫 联系我: haiboz8@163.com

  Contact me: haiboz8@163.com

• 💬 可以问我关于 AI Agent架构、推荐系统、大规模系统设计 的问题

  Ask me about AI Agent Architecture, Recommendation Systems, Large-scale System Design

Connect with me

Core Skills

AI & 机器学习

AI & Machine Learning

• 熟悉多Agent系统、langchain/langgraph/llamaindex、mem0、langfuse、熟悉（Leader Agent / Sub Agent、Plan-Task / ReAct / Reflexion / TOT）等编排和推理模式

  Proficient in multi-Agent systems, langchain/langgraph/llamaindex, mem0, langfuse, orchestration and reasoning patterns (Leader Agent / Sub Agent, Plan-Task / ReAct / Reflexion / TOT)

• 熟悉RAG检索召回（GraphRAG、BM25、Semantics Retrieval）

  Experienced in RAG retrieval and recall (GraphRAG, BM25, Semantics Retrieval)

• 熟悉模型数据集构建、微调和评估、模型在线Serving（vllm、TriTon Serving）

  Skilled in model dataset construction, fine-tuning, evaluation, and online serving (vLLM, Triton Serving)

• 熟悉大模型工具链（LiteLLM、reasoning mode、mcp/tools、context enginner）

  Familiar with LLM toolchain (LiteLLM, reasoning mode, MCP/tools, context engineering)

• 熟悉推荐系统（多路召回、精排模型、多目标优化、在线推理、系统工程）

  Expert in recommendation systems (multi-channel recall, ranking models, multi-objective optimization, online inference, system engineering)

后端开发

Backend Development

• 语言: Python, Java, Golang

  Languages: Python, Java, Golang

• 架构: 微服务架构、大规模分布式系统

  Architecture: Microservices architecture, large-scale distributed systems

• 基础设施: Kubernetes、CI/CD、DevOps

  Infrastructure: Kubernetes, CI/CD, DevOps

• 数据: Hadoop、Spark、Flink、Redis

  Data: Hadoop, Spark, Flink, Redis

前端开发

Frontend Development

• TypeScript, JavaScript, React, Vue

• 现代前端工程化

  Modern frontend engineering

Core Project Experience

Cultural Tourism AI Agent Platform

• 主导AI Agent平台技术架构和产品开发

  Led AI Agent platform technical architecture and product development

• 多Agent编排服务、模型推理服务、RAG检索召回服务

  Multi-Agent orchestration service, model inference service, RAG retrieval and recall service

• Memory记忆管理、Tools/MCP服务

  Memory management, Tools/MCP services

• 产品功能：语音智能问答、线路规划、数字人导游、AI拍照识景等

  Product features: Voice Q&A, route planning, digital human tour guide, AI photo recognition

ShuaBao APP Recommendation System (Jumei Youpin)

• 从0到1搭建推荐系统，支撑千万级用户

  Built recommendation system from scratch, supporting millions of users

• 用户人均观看时长提升100%（20分钟→55分钟）

  Increased average viewing time by 100% (20min → 55min)

• 系统规模：500台召回服务器、100台精排服务器、200+ Redis实例

  System scale: 500 recall servers, 100 ranking servers, 200+ Redis instances

• 数据规模：6PB，每日处理1TB数据

  Data scale: 6PB, processing 1TB daily

Tech Stack

GitHub Stats

Featured Projects

Tech Reading & Learning

• 📧 关注AI Newsletter和Blog

  Following AI Newsletters and Blogs

• 📄 喜欢读技术论文

  Love reading technical papers