你好，我是Hi, I'm

张嘉玮Jiawei Zhang · @chaw1

技术负责人 / 全栈架构师 / AI 工程化实践者 Tech Lead / Full-stack Architect / AI Engineering Practitioner

9 年软件开发与技术管理经验，精通 Java 与 Python 技术栈。擅长平台级架构设计、分布式高并发系统与 AI 工程化落地——最近一年从零到一主导了一个 AI 数据服务平台的建设，并把多智能体协作开发跑成了团队日常。 Nine years of software engineering and technical leadership across Java and Python. I design platform-level architectures, build high-concurrency distributed systems, and ship AI into production — most recently leading an AI data platform from zero to one and turning multi-agent development into the team's daily workflow.

看看我做过什么See my work → GitHub Email
杭州 · 达拉斯（中美两地）Hangzhou · Dallas (CN/US) Java · Python · Vue/TS · K8s M.S. 计算机科学，Marquette UniversityM.S. Computer Science, Marquette University 
jiawei@jiawis:~$ cat profile.yaml
role:      "Tech Lead / Full-stack Architect"
years:     9
languages: [Java, Python, TypeScript, Node.js]
focus:     [distributed-systems, AI-platforms, multi-agent-dev, CI/CD & GitOps]
now:       "SolarSense — AI data platform, 0 → 1, 21 repos, 20-person team"
status:    "open to interesting problems"   # 欢迎交流
9+年研发与技术管理years engineering & leadership
5,839近 12 个月代码提交commits in the last 12 months
21个代码仓库，横跨全栈repositories across the stack
400+次评审通过的 MR（20 人团队）MRs reviewed for a 20-person team
0.89s万级任务提交耗时（原 >2 min）for 10k-task submits (was >2 min)
01

现在在做What I'm building now

SolarSense AI 数据服务平台AI Data Platform

研发负责人 · 杭州景联文科技 · 2025.9 – 至今Head of Engineering · Hangzhou Jinglianwen Tech · Sep 2025 – present

从零主导整个平台的架构规划与技术选型：Java 多模块后端（数据集 / 治理 / 项目 / AI 能力 / 权限）+ Python 预处理服务，覆盖数据标注工具矩阵、能力平台与多租户治理域，支撑采集 → 标注 → 质检 → 审核 → 导出的全链路。

Led architecture and technology choices from a blank repo: a multi-module Java backend (datasets / governance / projects / AI capabilities / IAM) plus Python preprocessing services, covering an annotation tool matrix, a capability platform and multi-tenant governance — the full collect → annotate → QA → review → export pipeline.

  • 性能：批量任务处理优化，万级任务提交从 2 分钟以上降到 0.89 秒；故障注入验证容错（280 次重试场景验收）
  • 交付：ARM64 / 信创国产化环境私有化部署，客户现场实机验收；建立生产 / 预发 / 测试多环境发布与 CI/CD 门禁
  • 工程文化：制定 20 人团队的代码评审规范与合并纪律；落地基于 Claude Code / Codex 的多智能体协作开发（任务拆解 → 并行实现 → 独立验收）
  • 复杂场景：点云标注等高复杂度标注工具落地
  • Performance: batch task pipeline rebuilt — 10k-task submission from >2 min to 0.89 s; fault-injection tests covering 280 retry scenarios
  • Delivery: on-prem deployments on ARM64 / domestic (信创) stacks with on-site acceptance; prod / staging / test environments with CI/CD quality gates
  • Engineering culture: code-review and merge discipline for a 20-person team; multi-agent development with Claude Code / Codex (decompose → parallel implement → independent verification)
  • Hard problems: shipped high-complexity tools such as point-cloud annotation
stack Java · Spring Cloud · MySQL · Redis · Python · Vue3/TypeScript · Docker · K8s · GitLab CI/CD · LakeFS / Object Storage
提交（12 个月）commits (12 mo)5,839
非合并提交non-merge commits5,259
活跃天数active days243 / 365
代码仓库repositories21
发起 MR / 合并率MRs opened / merge rate163 / 80%
评审通过他人 MRMRs reviewed & approved400+
团队规模team size20
批量提交耗时bulk submit latency2 min → 0.89 s
02

技术栈Skills

// 语言与框架Languages & Frameworks

Java10y
Python8y
Spring Cloud / Boot7y
MyBatis8y
Vue.js / TypeScript5y
Netty5y
Node.js / Express3y
微信小程序WeChat Mini Programs4y

// 中间件与数据Middleware & Data

Redis7y
Kafka / RabbitMQ5y
MySQL / Oracle9y
MongoDB4y
Elastic Stack3y
分布式事务distributed tx 分布式锁distributed locks LakeFSpgvectorOSS / S3

// 架构与 AIArchitecture & AI

分布式系统设计Distributed systems7y
高并发优化High-concurrency tuning6y
Docker / K8s5y
AI 模型集成与调优AI model integration3y
多智能体协作开发Multi-agent development1y
微服务治理microservices CI/CD · GitOps LangChainClaude Code · Codex 信创 / ARM64ARM64 / domestic stacks
03

工作经历Experience

研发负责人Head of Engineering · 杭州景联文科技有限公司Hangzhou Jinglianwen Technology

2025.09 – 至今present
杭州Hangzhou, China

SolarSense AI 数据服务平台 — 从零主导多模块架构落地AI Data Platform — multi-module architecture from zero

  • 独立主导平台整体架构规划与技术选型，建设 Java 多模块后端与 Python 预处理服务，覆盖数据标注工具矩阵、能力平台与多租户治理域
  • 一年内主导代码提交 5,839 次，活跃 243/365 天，覆盖 21 个代码仓库；创建 MR 163 个（合并率 80%），审核团队 MR 逾 400 个
  • 批量任务处理性能优化：万级任务提交从 2 分钟以上降至 0.89 秒，并通过故障注入测试验证容错（280 次重试场景）
  • 数据全链路能力（采集 – 标注 – 质检 – 审核 – 导出）与点云标注等复杂场景落地
  • ARM64 / 信创国产化私有化部署与客户现场交付；建立 CI/CD 门禁与多环境发布流程；落地多智能体协作开发模式
  • Owned overall architecture and technology selection; built the multi-module Java backend and Python preprocessing services covering the annotation tool matrix, capability platform and multi-tenant governance
  • 5,839 commits in a year across 21 repositories, active 243/365 days; 163 MRs opened (80% merged), 400+ team MRs reviewed and approved
  • Batch task performance: 10k-task submission from >2 min to 0.89 s, validated with fault injection (280 retry scenarios)
  • End-to-end data pipeline (collect – annotate – QA – review – export) and complex tooling such as point-cloud annotation
  • ARM64 / domestic-stack on-prem deployments with on-site acceptance; CI/CD gates and multi-environment releases; multi-agent development workflow
stack Java · Spring Cloud · MySQL · Redis · Python · Vue3/TS · Docker · K8s · GitLab CI/CD · LakeFS

首席工程师Principal Engineer · JiawisTech Solutions LLC

2024.03 – 2025.05
美国德州 · 达拉斯Dallas, Texas, USA

AI 交互平台 — 多模型 AI 对话服务与企业级解决方案AI Interaction Platform — multi-model conversational AI for enterprises

  • 主导系统架构设计：微服务划分、Redis 缓存策略与 MongoDB 结构优化，输出并评审技术方案
  • 组织 Kafka 消息系统实现与 Docker 容器化部署，推动上线并提升服务可用性
  • 负责 AI 模型调优与场景集成，带队完成 LangChain 与 OpenAI 接口整合；推动"行为驱动提示"模块上线
  • Designed the system architecture — service boundaries, Redis caching strategy, MongoDB schema optimisation — and reviewed technical proposals
  • Drove the Kafka messaging layer and Docker-based deployment to production, improving availability
  • Owned model tuning and integration; led LangChain + OpenAI integration and shipped a behaviour-driven prompting module

电商优化 — Shopify 商店技术改进E-commerce — Shopify store engineering

  • 主导前端性能优化方案评审与多端加载提速策略
  • 带领 2 人小组完成 AI 客服系统部署与 NLP 问答流程梳理，显著提升客服响应率
  • 基于用户行为数据构建推荐引擎原型，负责核心算法逻辑与业务对接
  • Led front-end performance reviews and cross-device load-time improvements
  • Led a 2-person team deploying an AI customer-service system with an NLP Q&A flow, markedly improving response rates
  • Built a recommendation-engine prototype on user behaviour data, owning the core algorithm and business integration
stack Java · Python · Spring Boot/Cloud · Redis · MongoDB · Kafka · Docker · Nginx · OpenAI API · LangChain

软件工程师Software Engineer · 东方证券Orient Securities

2019.07 – 2021.01
上海Shanghai, China

证券交易系统重构 — C++ 系统的 Java 微服务化Trading system re-architecture — from C++ monolith to Java microservices

  • 参与核心交易模块 Java 重构，基于 Kafka 构建高性能消息系统，处理多系统海量交易数据
  • 基于 Netty 实现低延迟网络通信框架，毫秒级消息传输
  • 设计 Redis 分布式缓存与分布式锁，解决并发一致性，支撑每日百万级交易
  • Rebuilt core trading modules in Java with a Kafka-based high-throughput messaging layer across multiple upstream systems
  • Implemented a low-latency networking framework on Netty with millisecond-level delivery
  • Designed Redis-based distributed caching and locking for concurrency consistency at millions of trades per day

交易延迟监控 — 实时性能分析系统Trade-latency monitoring — real-time performance analytics

  • Python + Vue.js 开发毫秒级精度的延迟监控与瓶颈分析工具
  • 基于 ELK Stack 构建实时日志收集、延迟可视化与自动报警；开发性能异常检测算法
  • Built a millisecond-precision latency monitor and bottleneck analyser in Python + Vue.js
  • Real-time log pipeline, latency dashboards and alerting on the ELK Stack, plus an anomaly-detection algorithm for early risk signals
stack Java · Spring Cloud · Netty · Kafka · Oracle · Redis · Python · Vue.js · ELK

软件工程师Software Engineer · 联创集团Linkage Group

2017.09 – 2019.03
江苏南京Nanjing, China

徐州房地产交易平台 — 政府房产交易线上化Xuzhou real-estate transaction platform — digitising government property transactions

  • 带领 5 人团队，主导微服务架构设计、MySQL 索引优化与高并发事务处理方案
  • 输出 RabbitMQ 异步事务模型与 Redis 缓存策略；对接政府单位需求变更，组织阶段评审
  • Led a 5-person team on microservice design, MySQL index tuning and high-concurrency transaction handling
  • Delivered the RabbitMQ async-transaction model and Redis caching strategy; managed requirement changes with government stakeholders

"慧驾"汽车服务小程序"Huijia" automotive-services mini program

  • 独立负责小程序整体开发与迭代，设计 Node.js 中间层对接 Java 后端；建立 Code Review 与单元测试流程
  • Sole owner of the mini program end to end, with a Node.js middle layer in front of the Java backend; introduced code review and unit-testing practice
stack Java · Spring Boot · MyBatis · MySQL · RabbitMQ · Redis · Nginx · Activiti · Node.js · WeChat Mini Program
04

业余项目 & 实验Side projects & experiments

~/quanbot

QuanBot

自研量化交易系统：数据 → 上下文 → 判断 → 评分 → 预测 → 评估 → 报告的完整流水线，FastAPI 后端 + Web 前端，自托管运行。

A home-grown quantitative trading system: a full data → context → judge → score → forecast → evaluate → report pipeline, FastAPI backend with a web front end, self-hosted.

PythonFastAPINext.js
~/ai-infra

自托管 AI 基础设施Self-hosted AI infrastructure

在自己的云主机上跑一整套 AI 工作台：LobeChat（PostgreSQL / Redis / SearXNG / 对象存储）、向量数据库、Nginx + 自动证书，多智能体开发的日常试验田。

A full AI workbench on my own cloud host: LobeChat (PostgreSQL / Redis / SearXNG / object storage), a vector database, Nginx with automated TLS — my daily playground for multi-agent development.

DockerpgvectorNginx
~/netsec-lab

智能网络安全分析框架Intelligent network-security analytics

研究生课题：基于 Snort 3 + ELK 构建安全分析平台，集成 TensorFlow 做流量异常检测，在校园网边界部署传感器做实时监控。

Graduate research: a security analytics platform on Snort 3 + ELK with TensorFlow-based traffic anomaly detection, deployed as sensors at the campus network edge.

Snort 3ELKTensorFlow
05

教育背景Education

Marquette University · 美国马凯特大学Milwaukee, USA

计算机科学硕士 · GPA 3.7 / 4.0M.S. in Computer Science · GPA 3.7 / 4.0
2022 – 2023
联系我Get in touch

有意思的问题，随时聊Got an interesting problem? Let's talk.

架构咨询、AI 工程化落地、多智能体开发实践，或者只是想交流技术——邮件或 GitHub 都能找到我。Architecture, shipping AI to production, multi-agent development practice — or just talking shop. Email or GitHub both work.

zchaw127@gmail.com github.com/chaw1