Upload your resume. AI becomes your interviewer. Every question traces back to your actual experience.
上传简历，AI 化身大厂面试官，每道题都从你的真实经历出发。
Java Interview Coach is a Cowork skill that turns your resume into a personalized mock interview. Unlike generic question banks, every question is anchored to your actual projects and tech stack — the interviewer asks about your Redis usage, your SQL optimizations, your architectural choices.
| Feature | Description |
|---|---|
| Smart Resume Parsing | Extracts tech stack (tiered), projects, internships, and generates "chase points" — specific areas the interviewer will dig into |
| Three-Layer Deep Dive | Scene restoration → Technical decisions → Underlying principles. Goes deeper with each round |
| Resume-Anchored Questions | 80% of questions from your core tech stack, 20% from gaps you should fill. Every question starts with "I see on your resume…" |
| Per-Question Scoring | After each answer: score + what you nailed + what you missed + a model answer |
| Complete Retrospective | Dimension scores, per-question breakdown, resume optimization tips, and a 2-week study plan |
- Install: Drag
java-interview-coach.skillinto Cowork, or import via Settings - Upload: Drop your resume (PDF / DOCX / TXT / image) into the chat
- Trigger: Type
/java-interview-coachor say "帮我来个模拟面试" - Interview: AI parses your resume, then takes you through a 5-round interview (~40 min)
- Full Mock Interview — 5 rounds: intro, project deep dive, fundamentals, coding, reverse Q&A
- Project Deep Dive — Focused drilling on your projects only
- Knowledge Blitz — 10 fundamentals questions from your tech stack
- Single Question Explain — Drop a question, get a three-tier answer (Bronze / Gold / Diamond)
- University students preparing for campus recruitment (校招) in Java backend roles
- Junior developers (0–3 years) preparing for job switches
- Anyone who wants a structured, resume-driven interview simulation instead of random question banks
java-interview-coach/
├── SKILL.md ← Core instructions the AI follows
├── README.md ← This file
├── knowledge/
│ └── question-bank.md ← Reference question bank (8 domains + follow-up scripts)
└── examples/
└── sample-session.md ← Annotated walkthrough
Resume Upload → Parse tech stack & projects → User confirms
↓
Intro ← Project Deep Dive (3 layers) ← Fundamentals (5 angles) ← Coding ← Reverse Q&A
↓
Full Retro: scores + resume fixes + study plan
Found a question that's outdated, or want to add a tech domain? PRs welcome. The question bank (knowledge/question-bank.md) and the core instructions (SKILL.md) are both designed to be community-maintained.
MIT — use it, fork it, share it.
Java Interview Coach 是一个 Cowork 技能（Skill）。上传简历后，AI 化身大厂面试官，围绕你的真实项目经历和技术栈进行个性化模拟面试。跟随机抽题的面试机器人不同，这里的每一道题都能追溯到你简历中的具体内容——面试官会针对 你的 Redis 用法、你的 SQL 优化、你的 架构选型来追问。
| 功能 | 说明 |
|---|---|
| 简历智能解析 | 提取技术栈（三级分级）、项目经历、实习经历，自动生成"可追问点" |
| 三层深挖追问 | 场景还原 → 技术决策追问 → 底层原理深挖，层层递进 |
| 简历锚定出题 | 80% 从你的核心技术栈出题，20% 测知识盲区。每道题以"我看你简历上写了……"开场 |
| 逐题评分反馈 | 每题回答后给出：评分 + 亮点 + 遗漏点 + 满分话术 |
| 完整复盘报告 | 多维度评分、逐题分析、简历优化建议、两周突击学习路线 |
- 安装：把
java-interview-coach.skill文件拖入 Cowork，或在设置中导入 - 上传：把你的简历（PDF / DOCX / TXT / 图片）拖入对话
- 触发：输入
/java-interview-coach或者说"帮我做个模拟面试" - 面试：AI 解析简历后，带你走完 5 轮面试流程（约 40 分钟）
- 完整模拟面试 — 5 轮：自我介绍 → 项目深挖 → 八股文 → 手写代码 → 反问环节
- 项目深挖专项 — 只针对你的项目经历做深度追问
- 八股文突击 — 从你的技术栈出 10 道高频题
- 单题讲解 — 不上传简历，直接贴面试题得到三层回答（青铜/黄金/王者）
- Java 后端方向应届生（校招）
- 1–3 年经验的 Java 开发者（社招）
- 任何想要简历驱动、结构化面试模拟，而非随机抽题的开发者
java-interview-coach/
├── SKILL.md ← AI 执行的核心指令
├── README.md ← 本文件
├── knowledge/
│ └── question-bank.md ← 参考题库（8 大知识域 + 追问话术库）
└── examples/
└── sample-session.md ← 带注释的使用示例
上传简历 → 解析技术栈 & 项目 → 用户确认
↓
自我介绍 ← 项目深挖（3层追问）← 八股文（5个角度）← 手写代码 ← 反问
↓
完整复盘：逐题评分 + 简历优化 + 学习路线
面试题不是随机生成的。Skill 会扫描简历中的"信号词"来决定怎么问：
- 写了"优化了 SQL 性能" → 追问具体怎么优化的、explain 结果前后对比、索引怎么加的
- 写了"使用了 Redis" → 追问缓存更新策略、穿透/击穿/雪崩、为什么不用 Canal 同步
- 写了"负责 XX 模块" → 追问模块架构设计、数据流、最难的技术点
- 写了"高并发" → 追问实际 QPS、压测数据、如果扩大 10 倍能否扛住
八股文出题遵循五个角度轮换：核心原理 → 常见坑 → 版本差异 → 对比分析 → 真实故障排查。
Skill 会根据用户选择的目标公司类型调整面试策略：
- 大厂：基础八股 40% + 项目深挖 30% + 算法 30%，追问多且深，侧重底层原理
- 中厂：基础八股 50% + 项目深挖 40% + 偶尔算法，追问适中，侧重实用性
- 小厂：重项目经验，八股偏实用，较少算法
题库过时了？想补充新领域？欢迎提 PR。knowledge/question-bank.md 和 SKILL.md 都设计为社区可维护的结构。
MIT — 随意使用、Fork、分享给其他正在找工作的同学。