个人总结Professional Summary
AI 应用工程：AI Application Engineering: 6 年+ 前端与全栈经验，近 3 年专注 AI 产品落地。作为主要开发者主导 AI 内容创作 SaaS 端到端研发，个人代码贡献量居团队首位；自建 LLM 多模型路由层与鉴权失败自动降级；参与 Agent 平台二次开发，负责数据广场（数据集 → 知识库）从 0 到 1 与 MCP 工具接入。 6+ years in front-end and full-stack engineering, focused on AI products for the last 3 years. As the primary developer, led end-to-end R&D of an AI content-creation SaaS with the highest individual commit volume on the team; built an LLM multi-model routing layer with automatic fallback on auth failure; contributed to an Agent platform's secondary development, owning Data Square (dataset → knowledge base) from 0 to 1 and MCP tool integration.
端到端全栈交付：End-to-End Full-Stack Delivery: React / Next.js / TypeScript 前端 + Node / Go 服务端 + PostgreSQL / MySQL / Redis，独立完成 Docker 多容器（Web / Worker）部署与 CI/CD。近期独立承担微信小程序 + Go 后端（主导 Java → Go 重写并完成生产切流）与 React 管理台的开发与维护，覆盖支付、退款、库存全链路。 React / Next.js / TypeScript front-end with Node / Go services and PostgreSQL / MySQL / Redis; independently owns Docker multi-container (Web / Worker) deployment and CI/CD. Recently took sole charge of development and maintenance for a WeChat mini-program plus Go backend (leading a Java → Go rewrite through production cutover) and a React admin console, covering payment, refund, and inventory end to end.
AI 驱动的效率：AI-Driven Efficiency: 以 Cursor / Claude 为核心开发方式，配 Node / Python 自动化脚本。用 Babel 解析 AST 批量提取替换多语言文案，国际化周期缩短 1 周；优化 CRA / Webpack 构建，Rebuild 效率提升 30%+；推动 TypeScript 落地，核心页面 JS 报错率降至 0.02% 以下。 Cursor / Claude as the core development workflow, backed by Node / Python automation. Used Babel AST parsing to batch-extract and replace multilingual copy, cutting the i18n cycle by one week; optimized CRA / Webpack builds for 30%+ faster rebuilds; drove TypeScript adoption, reducing JS error rates on key pages below 0.02%.
工作经历Work Experience
前端开发工程师Front-End Engineer
- Agent 平台二次开发：Agent Platform Development: 基于 Bisheng 二次开发，负责数据广场从 0 到 1（数据集检索 / 分类递归筛选 / 详情 / 合成知识库与数据集转 QA 知识库导入）、租户管理（创建 / 列表 / 删除 / 批量注册）、MCP 工具接入与分组配置，以及品牌替换与登录页移动端适配。Secondary development on Bisheng: owned Data Square from 0 to 1 (dataset retrieval, recursive category filtering, detail view, knowledge-base synthesis and dataset-to-QA import), tenant management (create / list / delete / batch registration), MCP tool onboarding and grouping, plus branding replacement and mobile adaptation of the login page.
- AI 内容创作 SaaS（主要开发者）：AI Content-Creation SaaS (Primary Developer): 以 AI 辅助编程为核心开发方式，主导前后端端到端研发，个人提交量居团队首位；覆盖多平台内容生成（微信公众号 / 小红书 / 抖音 / B 站）、批量任务调度、Slate.js 富文本编辑器、AI 图像生成、品牌词库与知识库。Led end-to-end front-end and back-end R&D with AI-assisted coding as the core workflow, ranking first on the team by commit volume; covered multi-platform content generation (WeChat / Xiaohongshu / Douyin / Bilibili), batch task scheduling, a Slate.js rich-text editor, AI image generation, brand glossary, and knowledge base.
- LLM 调度与任务可靠性：LLM Scheduling & Task Reliability: 实现多模型路由层，支持流式输出与鉴权失败自动降级；基于 BullMQ 父子任务队列与 Redis 心跳 + 崩溃自动恢复，保障大批量生成任务稳定执行。Built a multi-model routing layer with streaming output and automatic fallback on auth failure; used BullMQ parent-child task queues with Redis heartbeat and crash auto-recovery to keep large-scale generation jobs running.
- 工程与交付：Engineering & Delivery: 设计维护 Docker 多容器架构（Web / Worker 独立容器），优化 Dockerfile 分层缓存；独立开发小红书小程序，打通 Web 端到小程序的发布链路。Designed and maintained a Docker multi-container architecture (isolated Web / Worker containers) with layered build caching; independently built the Xiaohongshu mini-program, completing the web-to-mini-program publishing pipeline.
前端开发工程师Front-End Engineer
- 核心产品迭代：Core Product Iteration: 主力参与 AI 社交产品（小冰岛 / X Eva / Chararu）前端研发，覆盖小程序、H5、App 内嵌模块与端内 Uniapp 热更包，累计服务用户超百万；个人在核心仓库提交 450+ 次。Key contributor to front-end R&D for AI social products (Xiaoice Island / X Eva / Chararu), covering mini-programs, H5, in-app modules and Uniapp hot-update bundles, serving over one million users; 450+ personal commits in the core repository.
- 多平台小程序矩阵：Multi-Platform Mini-Program Matrix: 负责小冰岛 / X Eva 系列多平台小程序（微信、抖音等）与端内 Uniapp 热更包开发，主导 4+ 小程序持续迭代与 10+ Web 页面，支持海外版（XevaMPUni.International）多端投放。Owned the Xiaoice Island / X Eva multi-platform mini-program family (WeChat, Douyin, etc.) and in-app Uniapp hot-update bundles; led continuous iteration of 4+ mini-programs and 10+ web pages, supporting the international build for overseas release.
- 运营后台全栈：Operations Console (Full-Stack): 独立承担运营后台前端与配套 Node API 开发，支撑活动配置与数据运营。Independently delivered both the operations-console front-end and its Node API, powering campaign configuration and data operations.
- 工程效能提升：Engineering Efficiency: 主导优化 Mac 平台打包插件，每次上线流程缩短 30%；为 Chararu 做国际化时，使用 Babel 解析源码 AST、批量提取并替换代码中的日文文案，编写自动化脚本实现多语言内容替换，国际化开发周期缩短 1 周。Led optimization of the Mac packaging plugin, shortening each release cycle by 30%; during Chararu's internationalization, used Babel to parse the source AST, batch-extract and replace Japanese copy, and wrote automation scripts for multilingual content replacement, cutting the i18n development cycle by one week.
- 跨平台能力：Cross-Platform Capability: 独立完成 H5_XEva 的 Apple 登录 + IAP 支付集成及阿拉伯语等多语言国际化适配，支持海外市场拓展。Independently completed Apple Sign-In + IAP payment integration and Arabic/multilingual i18n adaptation for H5_XEva, supporting overseas market expansion.
- 跨端协作：Cross-End Collaboration: 设计与客户端团队共建公用功能接口，标准化后类似功能开发效率提升 1 天+。Designed and co-built common function interfaces with the client team, standardizing similar features and improving development efficiency by 1+ day.
- Chrome 扩展探索：Chrome Extension Exploration: 内部技术探索项目，基于 Chrome Extension Manifest V3 + React 开发 AI 语音对话扩展，集成 Azure Cognitive Services SDK 实现实时 ASR 语音识别与 TTS 语音合成，结合 Live2D 虚拟人实时渲染，实现语音驱动的沉浸式 AI 社交体验。An internal tech-exploration project—developed an AI voice-conversation Chrome extension based on Manifest V3 + React, integrating the Azure Cognitive Services SDK for real-time ASR and TTS, combined with Live2D virtual avatar real-time rendering for an immersive, voice-driven AI social experience.
前端开发工程师Front-End Engineer
- 高流量 H5 开发：High-Traffic H5 Development: 负责映客 APP 内核心 H5 功能页与活动页开发与维护，名下 24 个业务 entry（贵族守护、盲盒、家族远征、PK 战绩、黑金俱乐部、集卡等），日 PV 总计超 80 万。Owned in-app H5 feature and campaign pages for the Inke app, with 24 business entries under my name (Guard, Blind Box, Family Expedition, PK Records, Black-Gold Club, Card Collection, etc.), totaling 800K+ daily PV.
- 跨端移动开发：Cross-End Mobile Development: 参与 React Native 跨平台项目（IKRNBundle），独立完成任务中心模块并实现 iOS/Android 双端复用，参与 RN 热更新与打包发布流程。Contributed to the React Native cross-platform project (IKRNBundle), independently delivering the task-center module with iOS/Android reuse, and working on RN hot-update and bundle release flows.
- 架构与体验优化：Architecture & Experience Optimization: 优化基于 CRA 的 React 项目架构与 Webpack 配置（多 entry 构建、预渲染、跨域代理），Rebuild 效率提升 30%+；沉淀通用组件库，相似需求开发效率提升 1 人日以上。Optimized the CRA-based React architecture and Webpack config (multi-entry builds, prerendering, dev-server proxying) for 30%+ faster rebuilds; built a shared component library, raising similar-requirement efficiency by 1+ person-day.
- 稳定性与 TypeScript 落地：Reliability & TypeScript Adoption: 从零搭建并推广团队 TS 开发环境，重构核心工具方法，核心页面 JS 报错率降至 0.02% 以下；负责支付业务前端（pay.busi.inke.com）并接入 ARMS 监控、过滤无效告警。Built and promoted the team's TS environment from scratch and refactored core utilities, reducing JS error rates on key pages below 0.02%; owned the payment front-end (pay.busi.inke.com) with ARMS monitoring and alert-noise filtering.
项目经历Project Experience
已上线（尚未推广）· 由「滑雪去」发起，本人负责全栈开发与维护Launched (pre-promotion) · initiated by the “Ski” org; I own full-stack development and maintenance
技术栈：Tech Stack: 微信小程序（原生 + TypeScript）· Go（chi / GORM / MySQL）· React 19 + TanStack Router · Docker · PlaywrightWeChat Mini-Program (native + TypeScript) · Go (chi / GORM / MySQL) · React 19 + TanStack Router · Docker · Playwright
- 由「滑雪去」发起的滑雪场景产品（拼车 / 拼房 / 二手雪票三条业务线），已上线运行；本人独立承担小程序端 + Go 后端 + React 管理台的开发与持续维护。A ski-scenario product launched by the “Ski” org (carpool / room-share / second-hand ticket lines), live in production; I independently own development and ongoing maintenance of the mini-program, Go backend, and React admin console.
- Java → Go 重写并生产切流：Java → Go Rewrite with Production Cutover: 主导后端从 Java（SpringMVC + MyBatis）重写为 Go，完成等价性验证与线上切流，旧服务观察期保留可回滚。Led the backend rewrite from Java (SpringMVC + MyBatis) to Go, completing equivalence verification and production cutover with the legacy service kept for rollback.
- 交易与资金正确性：Transactional & Payment Correctness: 实现微信支付 / 退款全链路、商城库存占用与释放、优惠券回滚与过期关单，核心路径用条件状态迁移 + 事务保证幂等，并补齐审计与 CSRF 防护。Built WeChat payment and refund end to end, mall inventory reservation/release, coupon rollback and order expiry, using conditional state transitions plus transactions for idempotency, with audit logging and CSRF protection.
- 工程质量：Engineering Quality: 建立 ADR 决策记录与数据库迁移体系，管理台 240+ 测试 + Playwright E2E，接入 OpenTelemetry 链路追踪，镜像以 non-root 运行并通过漏洞扫描。Established ADR decision records and a DB migration system; 240+ console tests plus Playwright E2E, OpenTelemetry tracing, and non-root images passing vulnerability scans.
技术栈:Tech Stack: Next.js / Node.js / TypeScript / PostgreSQL / Redis / BullMQ / Drizzle ORM / Vercel AI SDK
- 作为主要开发者主导全平台 AI 内容创作系统前后端端到端研发，个人提交量居团队首位（611+ commits）。As the primary developer, led end-to-end front-end and back-end R&D of the platform-wide AI content-creation system, ranking first on the team by commit volume (611+ commits).
- 前端与富交互:Front-End & Rich Interaction: 构建基于 Slate.js 的富文本编辑器，集成 AI 辅助改写、悬浮工具栏、图片本地上传与拖拽排序等功能。Built a Slate.js-based rich-text editor with AI-assisted rewriting, a floating toolbar, local image upload, and drag-and-drop sorting.
- 服务端与模型调度:Server & Model Scheduling: 实现 LLM 多模型路由层，统一接入多家模型，支持流式输出与鉴权失败自动降级（401/403 识别 + 冷却窗口），以及多平台并行 Artifact 状态管理。Built an LLM multi-model routing layer unifying multiple providers, with streaming output and automatic fallback on auth failure (401/403 detection with a cooldown window), plus multi-platform parallel Artifact state management.
- 后端任务中心设计:Back-End Task-Center Design: 开发批量内容生成系统，基于 BullMQ 父子任务队列调度，结合 Redis 心跳监测与崩溃自动恢复，保障大批量任务稳定执行；并实现定时发布队列与 GEO 监测 Worker。Built a batch content-generation system on BullMQ parent-child task queues with Redis heartbeat and crash auto-recovery for large-scale jobs; also delivered scheduled-publishing and GEO-monitoring workers.
技术栈：Tech Stack: React / Vue 3 / TypeScript / Uniapp / 小程序多平台
- 核心贡献者（2年），主导 Chararu 多个核心模块：日记系统、游戏房间、私密房间、小说游戏、角色对话系统、通知与本地化配置。Key contributor (2 years), leading multiple core modules of Chararu: diary system, game rooms, private rooms, novel game, character dialogue system, notifications, and localization config.
- 负责小冰岛 / X Eva 多平台小程序与端内 Uniapp 热更包，主导 4+ 小程序持续迭代与 10+ Web 页面开发，保障高并发下的用户体验。Owned the Xiaoice Island / X Eva multi-platform mini-programs and in-app Uniapp hot-update bundles; led continuous iteration of 4+ mini-programs and 10+ web pages, ensuring UX under high concurrency.
- 完成 H5_XEva 的 Apple 登录、IAP 支付及阿拉伯语等多语言国际化适配，支持海外市场拓展。Completed Apple Sign-In, IAP payment, and Arabic/multilingual i18n adaptation for H5_XEva, supporting overseas expansion.
技术栈：Tech Stack: React / TypeScript / React Native / Redux / Webpack / ARMS
- 独立负责 24 个业务 entry 的高并发 H5 功能页（日 PV 80 万+）开发与维护，覆盖贵族守护、盲盒、家族、PK、集卡等核心玩法。Independently owned 24 business entries of high-concurrency H5 pages (800K+ daily PV), covering Guard, Blind Box, Family, PK, and card-collection gameplay.
- 参与 React Native 跨平台项目（IKRNBundle），独立完成任务中心功能模块，实现 iOS/Android 双端复用与热更发布。Contributed to the React Native cross-platform project (IKRNBundle), independently delivering the task-center module with iOS/Android reuse and hot-update release.
- 推动 TS 落地，核心页面 JS 报错率降至 0.02% 以下；优化 Webpack 多 entry 构建与预渲染，Rebuild 效率提升 30%+。Drove TS adoption, reducing JS error rate on key pages below 0.02%; optimized Webpack multi-entry builds and prerendering for 30%+ faster rebuilds.
- 个人独立开发并发布至 VS Code Marketplace 的 VS Code 扩展插件，累计下载量 746+。A VS Code extension independently developed and published to the VS Code Marketplace, with 746+ cumulative downloads.
- 提供快速插入/删除 console.log、console.error 等调试语句的命令与快捷键，优化日常前端调试工作流。Provides commands and shortcuts to quickly insert/remove debugging statements such as console.log and console.error, optimizing the daily front-end debugging workflow.
技术栈Technical Skills
前端框架：Front-End Frameworks: React 18 / Next.js / Vue 3 / Uniapp / Tailwind CSSReact 18 / Next.js / Vue 3 / Uniapp / Tailwind CSS
跨端开发：Cross-End Development: React Native / 微信小程序 / 抖音小程序 / Uniapp 热更 / Chrome 扩展React Native / WeChat Mini-Program / Douyin Mini-Program / Uniapp hot-update / Chrome Extension
工程化：Engineering: TypeScript / Webpack 5 / Vite / Docker / CI-CD / Playwright / OpenTelemetryTypeScript / Webpack 5 / Vite / Docker / CI-CD / Playwright / OpenTelemetry
后端/数据：Back-End / Data: Node.js / Go / PostgreSQL / MySQL / Redis / GORM / Drizzle ORM / BullMQNode.js / Go / PostgreSQL / MySQL / Redis / GORM / Drizzle ORM / BullMQ
AI 工具链：AI Toolchain: Cursor / Claude（AI 辅助编程）· Vercel AI SDK / LLM 多模型路由与自动降级 / MCP 工具集成 / RAG 知识库 / Azure Cognitive ServicesCursor / Claude (AI-assisted coding) · Vercel AI SDK / LLM multi-model routing & fallback / MCP tool integration / RAG knowledge base / Azure Cognitive Services
教育经历Education
统招全日制本科Full-time bachelor's degree (national unified enrollment)