FDE CAREER TRACK

FDE 职业进阶路径FDE Career Track

这一组课程不是刷题营,也不是照搬某家公司面试经验,而是把公开的 FDE 面试准备资料整理成可训练、可交付、可复盘的职业能力路径。This track does not copy any company interview process. It turns public FDE preparation notes into trainable career capabilities for delivery, validation, and retrospectives.

POSITIONING

FDE FAN 的核心不是让学员死磕语法,而是训练学员用 Codex、Claude Code、Cursor、Antigravity 等 AI 工具完成真实交付。Google 文件夹里的材料适合转化为“高阶职业进阶”模块:面试只是表层,底层是意图拆解、Agent 系统判断、技术表达和项目复盘。 FDE FAN is not about grinding syntax. It trains learners to use AI tools such as Codex, Claude Code, Cursor, and Antigravity for real delivery. The Google folder becomes a higher-level career track: interviews are the surface; intent framing, agent design, technical communication, and retrospectives are the core.

模块结构

MAP

FDE 职业能力地图FDE Career Capability Map

把面试准备改写成职业能力准备:角色动机、项目表达、技术判断、协作沟通和复盘能力。Reframe interview prep as career readiness: role motivation, project storytelling, technical judgment, collaboration, and retrospectives.

VIBE

AI 工具协作开发AI-Assisted Vibe Coding

从逐行写代码转向描述意图、拆任务、审输出、跑验证,用 Codex、Claude Code、Cursor 和 Antigravity 交付 Demo。Move from writing every line to framing intent, decomposing tasks, reviewing output, and validating demos with AI coding tools.

AGENT

生产级 Agent 系统设计Production Agentic System Design

学习编排、结构化输出、上下文压缩、RAG、工具调用、护栏、成本控制和系统化评估。Study orchestration, structured output, context compression, RAG, tool calls, guardrails, cost control, and systematic evaluation.

DSA

DSA Lite 技术表达DSA Lite for Technical Communication

保留算法模式训练,但不把 FDE FAN 做成刷题站:目标是能讲清复杂度、边界条件和方案权衡。Keep algorithm patterns as communication training, not a problem-grinding track: explain complexity, edge cases, and tradeoffs.

MOCK

Mock Interview 与项目复盘Mock Interview and Project Review

用真实项目做 30 分钟演示、Prompt + 输出审查、失败诊断和 STAR 复盘。Use real projects for 30-minute demos, prompt-output review, failure diagnosis, and STAR retrospectives.

与主课程的关系

基础课程学会指挥 AI 工具理解项目结构、需求拆解、Prompt 到页面、Agent 工作流和部署交付。
行业课程进入真实业务把跨境电商、文旅、企业诊断等场景拆成可交付 Demo、流程和文档。
职业进阶证明可被信任用 Vibe Coding、Agent 系统设计、DSA Lite 和 Mock Review 证明自己能处理模糊需求、错误输出和上线边界。

参考来源