Skip to content

README.md

Latest commit

 

History

History
159 lines (118 loc) · 4.48 KB

README.md

File metadata and controls

159 lines (118 loc) · 4.48 KB

🎯 Virtual Interview Navigator

An AI-powered mock interview platform that generates personalized questions from your CV and delivers structured performance feedback.

Status Backend Frontend AI License

📖 Overview

Virtual Interview Navigator lets you upload your CV, specify a target job role and description, and step into a full mock interview — all powered by OpenAI. The system tailors every question to your specific background and grades your answers with detailed, actionable feedback.

✨ Features

Feature Description
🧠 Personalized Questions CV + Job Description fed directly into GPT-4o-mini for context-aware question generation
🎙️ Voice Input Speak your answers using the native Browser Speech Recognition API
📊 Scored Feedback Per-question scores (0–1.0) with Correct / Partial / Wrong badges
📝 Full Report Markdown report covering Communication, Confidence, Technical Accuracy & Improvement Areas
🔐 Secure Auth Stateless JWT authentication with protected routes
📄 Document Parsing Supports PDF and DOCX CV uploads

🛠️ Tech Stack

Frontend

  • React 18 (Vite) — component-based UI
  • Tailwind CSS — utility-first responsive styling
  • Axios — HTTP client with JWT interceptors
  • Web Speech API — browser-native voice recognition
  • React Markdown — renders AI-generated feedback reports

Backend

  • Python / Flask — lightweight REST API
  • PostgreSQL (Neon) — relational data persistence
  • SQLAlchemy — ORM and migration management
  • OpenAI GPT-4o-mini — question generation & answer analysis
  • PyPDF2 / python-docx — CV text extraction

Infrastructure

  • Vercel — frontend deployment
  • Render — backend container hosting
  • Neon — serverless PostgreSQL

🗂️ Project Structure

interviewnav/
├── backend/
│   ├── app.py              # Flask app, all API routes
│   ├── model.py            # SQLAlchemy models
│   ├── prompts.py          # OpenAI prompt templates
│   ├── config.py           # Environment configuration
│   ├── extensions.py       # DB & JWT initialization
│   └── requirements.txt
└── frontend/
    ├── src/
    │   ├── pages/          # Home, Login, Interview, Report …
    │   ├── services/       # api.js, authService.js
    │   └── App.jsx
    └── package.json

⚙️ Local Setup

Prerequisites

  • Python 3.11+
  • Node.js 18+
  • PostgreSQL (local or Neon)

1. Database

CREATE DATABASE "InterviewNavigator";

2. Backend

cd backend
python -m venv venv

# Activate (choose your OS)
# Windows:   venv\Scripts\activate
# Mac/Linux: source venv/bin/activate

pip install -r requirements.txt
flask db upgrade
python app.py

Create a backend/.env file:

DATABASE_URL=postgresql://user:password@localhost:5432/InterviewNavigator
OPENAI_API_KEY=sk-...
SECRET_KEY=your-secret-key
CORS_ORIGINS=http://localhost:5173

3. Frontend

cd frontend
npm install
npm run dev

Create a frontend/.env file:

VITE_API_URL=http://localhost:5000

🚀 Deployment

Backend — Render

Setting Value
Build Command ./build.sh
Start Command gunicorn app:app --timeout 300
Env Vars DATABASE_URL, OPENAI_API_KEY, SECRET_KEY, CORS_ORIGINS

Frontend — Vercel

Setting Value
Framework React (Vite)
Env Vars VITE_API_URL → your Render backend URL (no trailing slash)

🔄 How It Works

1. Register / Login  →  JWT issued
2. Upload CV (PDF/DOCX)  →  Text extracted & stored
3. Enter Job Role + JD  →  Sent to GPT-4o-mini with CV context
4. Answer questions  →  Via text or voice input
5. Complete interview  →  GPT analyses all responses
6. View Report  →  Scores, badges & improvement suggestions