Intreview - AI-Powered Interview Practice Platform

🚀 Overview

Intreview is an advanced AI-powered platform that helps users practice and improve their interview skills through real-time analysis of speech patterns, body language, and overall presentation. Built for FinHacks, this project combines cutting-edge AI technologies to provide comprehensive feedback on interview performance.

🌟 Key Features

• Real-time Speech Analysis: Advanced speech pattern recognition with filler word detection
• Computer Vision Feedback: Posture monitoring, eye contact tracking, and sentiment analysis
• Live Performance Metrics: Instant feedback on speaking pace, clarity, and engagement
• Comprehensive Results: Detailed post-interview analysis with actionable recommendations
• Professional UI/UX: Clean, modern interface optimized for interview practice

🏗️ Architecture

Backend (FastAPI + Python)

• Speech Processing: AssemblyAI integration for advanced speech analytics
• Computer Vision: MediaPipe for real-time facial landmark detection
• Real-time Communication: WebSocket connections for live feedback
• Data Storage: MongoDB for session and analysis data
• Authentication: JWT-based secure user authentication

Frontend (React + TypeScript)

• Modern UI: Tailwind CSS with custom styling
• Real-time Updates: WebSocket integration for live feedback display
• Interactive Components: Dynamic progress bars and real-time metrics
• Results Visualization: Comprehensive analytics dashboard

🛠️ Technology Stack

Component	Technology	Purpose
Backend API	FastAPI + Python	High-performance async API
Frontend	React + TypeScript	Modern, type-safe UI
Speech AI	AssemblyAI	Advanced speech analysis
Computer Vision	MediaPipe + OpenCV	Real-time video processing
Database	MongoDB	Document storage
Styling	Tailwind CSS	Utility-first CSS framework
Real-time	WebSockets	Live communication
Authentication	JWT + bcrypt	Secure user management

📋 Prerequisites

• Python 3.11+
• Node.js 18+
• MongoDB (local or cloud instance)
• AssemblyAI API Key (Get one here)
• Webcam and Microphone for interview practice

🎯 Core Features

Real-time Speech Analysis

• Words Per Minute: Optimal speaking pace tracking (target: ~150 WPM)
• Filler Word Detection: Identifies "um", "like", "you know", etc.
• Speech Intelligibility: AI-powered clarity scoring
• Confidence Scoring: Word-level confidence analysis

Computer Vision Analytics

• Posture Monitoring: Real-time posture assessment
• Eye Contact Tracking: Measures attention and engagement
• Sentiment Analysis: Detects positive, neutral, or negative expressions
• Facial Landmark Detection: 468-point facial mesh analysis

Performance Metrics

• Overall Performance Score: Weighted combination of all metrics
• Speech Analysis: 40% intelligibility, 30% filler words, 30% pace
• Visual Analysis: Eye contact, posture, and sentiment scoring
• Key Moments: Timestamped highlights and areas for improvement

📊 API Endpoints

Authentication

POST /auth/register    - User registration
POST /auth/token      - Login and token generation
GET  /auth/me         - Get current user info

Session Management

POST /api/sessions/start                    - Start new interview session
POST /api/sessions/{id}/end                 - End session with analysis
GET  /api/sessions/{id}/analysis           - Get session analysis
GET  /api/sessions/history                 - Get user's session history

Real-time Analysis

WebSocket /api/ws/video                    - Real-time video processing
POST /analysis/speech                      - Speech analysis endpoint

🔧 Development

Project Structure

intreview/
├── backend/
│   ├── app/
│   │   ├── api/routes/          # API route handlers
│   │   ├── services/            # Business logic
│   │   ├── db/models/          # Database models
│   │   └── core/               # Configuration
│   ├── main.py                 # FastAPI application
│   └── requirements.txt
├── frontend/
│   ├── src/
│   │   ├── components/         # Reusable components
│   │   ├── pages/             # Main page components
│   │   ├── contexts/          # React contexts
│   │   └── data/              # Static data
│   ├── public/                # Static assets
│   └── package.json
└── README.md

🎨 UI Components

Camera Interface

• Live video feed with real-time feedback
• Recording controls and session management
• Progress indicators and metric displays
• Interview question rotation

Results Dashboard

• Comprehensive performance analysis
• Interactive charts and progress bars
• Video playback with timestamped moments
• Downloadable PDF reports
• Actionable improvement recommendations

🔍 Analysis Algorithms

Speech Intelligence Scoring

speech_intelligibility = (
    weighted_confidence * 0.6 +    # Base confidence
    filler_word_score * 0.4        # Filler penalty
)

Filler Word Scoring

• Excellent: ≤1 filler per minute (100%)
• Good: ≤2 fillers per minute (80%)
• Fair: ≤4 fillers per minute (60%)
• Poor: >4 fillers per minute (exponential decay)

Visual Analysis

• Eye Contact: Based on face position relative to center
• Posture: Facial landmark positioning analysis
• Sentiment: Mouth curvature and facial expression detection

🏆 Hackathon Achievement

Built for FinHacks 2025 - Recognized for innovation in AI-powered interview preparation technology.

👥 Team

• David Yu - Full-stack development, AI integration
• Bew - Frontend development, UI/UX design