✨ PrepMind

A next-generation, AI-powered interview preparation and tailored resume generation platform built to bridge the gap between candidates and job descriptions.

---

🔮 Vision

PrepMind aims to democratize competitive job-hunting by providing job seekers with bespoke prep environments. By extracting profile data directly from uploaded PDF resumes or manual self-descriptions, PrepMind uses advanced generative AI to align candidate capabilities with specific job descriptions—formulating interactive match scores, structured learning roadmaps, targeted interview prep, and keyword-optimized resume generation to bypass ATS filter grids.

---

🎭 Problem & Solution

Problem	PrepMind's Solution
Candidacy Ambiguity	Generates a metric-driven 0-100 Match Score by comparing the candidate's history side-by-side with the target role.
Hidden Skill Gaps	Pinpoints exact missing technologies and methodologies, tagging them as High, Medium, or Low severity checklists.
Generic Interview Prep	Engineers custom Technical & Behavioral questions unique to the candidate and role, providing interviewer intention and suggested answers.
Unstructured Study Paths	Computes a chronological, Day-by-Day Preparation Roadmap complete with focus topics and concrete tasks.
ATS Rejections	Compiles job-specific, keyword-rich resumes dynamically as polished, A4 downloadable PDFs.

---

🚀 Core Features & Capabilities

Capability	Description
📊 Match & Gap Analytics	Provides an analytical baseline representing role-fit and displays severity badges indicating which lacking skills to study first.
🧠 Deep Prep Questions	Renders realistic behavioral scenarios (following the STAR framework) and core coding queries customized to the target stack.
🗺️ Interactive Study Map	Houses chronological day-cards showing specific tasks. Includes checklist items and progressive learning blocks.
📄 ATS-Tailored Resumes	Generates professionally styled HTML resumes tailored to the job's keywords and converts them on-demand to PDF.
🔒 Secure Credentials	Features robust JWT session handling via HTTP-only cookies and database-backed token blacklisting for safe logouts.
📁 Plan History Dashboard	Integrates an elegant, real-time search and filter panel allowing users to instantly manage previous preparation reports.

---

🏗️ Platform Architecture & Data Flow

PrepMind utilizes a modern full-stack decoupled architecture. Below is a map outlining the code dependencies and communication flows:

• Front-End Client: Engineered as a high-speed React 19 SPA compiled via Vite, using Tailwind CSS v4 and oklch-based dark-glassmorphism accents.
• Back-End API: Formulated as an Express.js Node API connected to MongoDB Atlas. Integrates the new Google GenAI SDK (gemini-2.5-flash) utilizing structured JSON schema modes.
• Dynamic Document Renderer: Integrates a hybrid browser engine. Launches Puppeteer locally for lightweight development, and automatically maps @sparticuz/chromium + puppeteer-core in production for flawless Vercel serverless execution.

---

📂 Codebase Project Structure

PrepMind/
├── client/                      # Vite + React Frontend Client
│   ├── src/
│   │   ├── components/
│   │   │   └── ui/              # Low-level layout components (Card, Dialogue, Inputs)
│   │   ├── features/
│   │   │   ├── auth/            # Auth pages (Login, Register), context bounds & hooks
│   │   │   └── interview/       # Dashboards, preparation lists, and sidebar modules
│   │   ├── App.tsx              # Main entry point mounting hooks & context wrappers
│   │   ├── index.css            # Tailwind resets, OKLCH theme presets, and styles
│   │   └── app.routes.tsx       # React Router 7 mappings and protected views
│   └── package.json
│
└── server/                      # Express.js backend API server
    ├── config/                  # Database connections
    ├── routes/                  # REST route bindings (Auth, Interviews)
    ├── controllers/             # Business logic handlers (PDF extractors, reports)
    ├── models/                  # MongoDB schemas (Users, Reports, Blacklist)
    ├── services/                # Gemini integrations & Puppeteer engines
    ├── middlewares/             # Authorization bounds and Multer file upload filters
    ├── server.js                # Port bindings & serverless handler endpoints
    └── package.json

---

💻 Technical Stack

Category	Technologies
Frontend Framework	React, TypeScript, Vite, React Router
State & Fetching	React Hooks, Context API, Axios Client
UI & Styling	Tailwind CSS, SCSS/SASS, shadcn/ui, Lucide Icons
Backend & Logic	Node.js, Express.js (ES Modules)
Database	MongoDB Atlas, Mongoose ODM
AI Integration	Google Generative AI (@google/genai) with JSON response schemas
Document Processing	Puppeteer Core, @sparticuz/chromium, pdf-parse (mehmet-kozan/pdf-parse)
Security	JSON Web Tokens (JWT), HTTP-Only Cookies, bcryptjs, Token Blacklisting

---

🚀 Getting Started

1. Setup Environment Variables

Create a .env file in the server root directory:

PORT=3000
MONGODB_URI=mongodb+srv://<username>:<password>@cluster.hn75bvj.mongodb.net/prep-mind
JWT_SECRET=your_jwt_cryptographic_secret_key
GEMINI_API_KEY=your_google_gemini_api_key
NODE_ENV=development

2. Install Dependencies

Install packages for both modules from the root directory:

# Install Server dependencies
cd server
npm install

# Install Client dependencies
cd ../client
npm install

3. Run the Applications

Start the development servers concurrently:

Terminal 1 — Backend API Server

cd server
npm run dev

Terminal 2 — Frontend Client

cd client
npm run dev

4. Local Access Points

• Frontend Client: http://localhost:5173
• Backend API Service: http://localhost:3000/api