Virtual Interview Navigator – AI Interview System

Virtual Interview Navigator is an AI-powered full-stack application designed to simulate real-world technical interviews. It generates personalized interview questions based on the candidate's CV, Job Role, and Job Description, and provides comprehensive, structured feedback on performance.

🚀 Key Features

🧠 Intelligent Orchestration

• End-to-End Request Pipeline: Validates, persists, and processes interview context (CV, Role, JD) to generate highly relevant questions.
• RAG-Lite Personalization: Uses "Retrieval-Augmented Generation" principles to tailor prompts based on uploaded documents.

🎙️ Interactive Interface

• Voice-Enabled: Integrated Browser Speech Recognition allows candidates to speak their answers hands-free.
• Real-Time Responsiveness: Mobile-friendly, adaptive UI centered around user experience.

📊 Comprehensive Feedback System

• Detailed Scoring: Assigns weighted scores (0-1.0) per question based on relevance and correctness.
• Structured Analysis: Generates rich Markdown reports analyzing:
  • Communication Skills
  • Confidence
  • Technical Accuracy
  • Areas for Improvement
• Visual Badges: Instant "Correct", "Partial", or "Wrong" status indicators.

🏗️ Robust Architecture

• Stateless Backend: Flask REST API with JWT-based authentication and PostgreSQL persistence.
• Document Ingestion: Robust PyPDF2 and python-docx pipelines for text extraction.
• AI Integration Layer: Isolated service layer ensuring clean separation between business logic and LLM inference.

---

🛠️ Tech Stack

Frontend

• React.js (Vite): High-performance UI library.
• Tailwind CSS: Utility-first styling for responsive design.
• Web Speech API: Native browser speech recognition.
• React Markdown: Rich text rendering for AI reports.
• Axios: HTTP client with interceptors for JWT handling.

Backend

• Python Flask: Lightweight, extensible web framework.
• PostgreSQL (Neon DB): Relational database for persistent storage.
• SQLAlchemy ORM: Database abstraction layer.
• OpenAI GPT-4o: Large Language Model for inference.
• PyPDF2 / python-docx: Document processing.

Infrastructure

• Render: Backend container orchestration.
• Vercel: Frontend edge deployment.
• Neon: Serverless PostgreSQL.

---

🔧 Local Setup

Prerequisites

• Python 3.11+
• Node.js 18+
• PostgreSQL

1. Database

Create a database named InterviewNavigator.

CREATE DATABASE "InterviewNavigator";

2. Backend

cd backend
python -m venv venv
# Windows: venv\Scripts\activate | Mac/Linux: source venv/bin/activate
pip install -r requirements.txt
flask db upgrade
python app.py

Note: Ensure .env is configured with DATABASE_URL and OPENAI_API_KEY.

3. Frontend

cd frontend
npm install
npm run dev

Note: Ensure .env has VITE_API_URL=http://localhost:5000 locally.

---

🚀 Deployment Guide

Backend (Render)

1. Build Command: ./build.sh (Installs deps + runs DB migrations).
2. Start Command: gunicorn app:app --timeout 300 (Increased timeout for long AI generation).
3. Env Vars: DATABASE_URL (Neon), OPENAI_API_KEY, CORS_ORIGINS (Vercel URL).

Frontend (Vercel)

1. Framework: React (Vite).
2. Env Vars: VITE_API_URL (Your Render Backend URL, no trailing slash).

---

🛡️ License

MIT License.