How stephenthangaraja/resume-builder works
Project type and purpose
• This is a full-stack web application: a Resume Builder with ATS (Applicant Tracking System) and relevance-scoring features, AI-driven suggestions, and cover-letter generation.
• Frontend: React + Vite + Tailwind CSS SPA.
• Backend: Node.js + Express server.
• Persistence & hosting: Firebase (Auth, Firestore, Storage, Hosting).
• AI: Fine-tuned Gemini 1.5 Flash for suggestions and scoring.
File structure
• Top-level “frontend” and “backend” directories.
• Frontend substructure: “public/”, “src/” with “pages/”, “components/”, “contexts/”, “assets/”, “lib/”. This is a component-based SPA using Context API for state.
• Backend: single Express entrypoint (server.js), package.json, an uploads folder for file storage.
• Config files: Vite (frontend), Tailwind, ESLint, PostCSS, jsconfig, dotenv (for backend).
• Firebase config lives alongside src; the React app likely uses the Firebase SDK directly for Auth and Firestore.
Extract architecture insights from the README
• Frontend communicates with Firebase SDK for authentication, Firestore reads/writes, and Storage for file assets.
• Backend exposes RESTful endpoints for ATS score evaluation, relevance scoring, cover-letter generation, file uploads, and possibly for AI model integration.
• Backend must also coordinate calls to the Gemini 1.5 AI model (via API).
• Hosting: static build on Firebase Hosting or similar; Node server deployed separately (e.g. on Heroku, GCP, or Firebase Functions).
a. Main components
• Client (Browser) running the React SPA
• Firebase Services
• Authentication
• Firestore Database
• Storage
• (Optionally) Hosting
• Express Backend API
• AI Model Service (Gemini 1.5 Flash)
• Static Assets / File Upload (uploads folder)
b. Relationships and interactions
1. User → React SPA (UI pages, forms, resume templates)
2. React SPA ↔ Firebase Auth (signup, login)
3. React SPA ↔ Firestore (CRUD resume data)
4. React SPA ↔ Firebase Storage (upload/download resumes, cover letters)
5. React SPA → Backend API (HTTP REST calls) for:
• ATS scoring
• Relevance scoring
• AI-powered suggestions
• Cover-letter generation
6. Backend API → Gemini 1.5 AI Service (RPC or HTTP)
7. Backend API → Firestore/Storage (optionally store analysis results)
c. Architectural patterns & principles
• Component-based UI (React) with Context API for global state.
• SPA served statically; API layer separated from UI.
• RESTful service design in Express.
• Eventual consistency: UI fetches async results (analysis, scores).
• Single-responsibility principle: Backend handles compute-heavy AI tasks, frontend handles UX.
d. Key technologies
• Frontend: React, Vite, Tailwind CSS, Firebase JS SDK
• Backend: Node.js, Express, Firebase Admin SDK (optional), dotenv
• Database & Storage: Firebase Firestore, Firebase Storage
• AI Service: Gemini 1.5 Flash (external API)
Detail
• Asynchronous flows (e.g., user triggers analysis → API processes → UI refreshes).
• Error-handling or retry paths if relevant (e.g., failed AI call).