CodeSage: AI Technical Interviewer

Building Intelligence That Evaluates Intelligence

CodeSage is an AI-powered technical interviewer that conducts live, adaptive coding interviews with human-like insight. Built for the Eightfold AI × ArIES Hackathon.

🚀 Features

1. Real-time Code Analysis Engine

• Live Code Tracking: Continuously monitors code as it's written
• Judge0 Integration: Secure code execution using industry-standard Judge0 API
• Multi-language Support: Execute code in 15+ programming languages
• Instant Error Detection: Real-time syntax and runtime error identification
• Performance Metrics: Execution time, memory usage, and Big O complexity analysis
• Test Case Validation: Automatic validation against predefined test cases

2. Agentic Interviewer System

• Dynamic Difficulty: Adaptive problem complexity based on performance
• Progressive Hints: Three-tier hint system (Nudge → Guide → Direction)
• Context-Aware Questions: Intelligent follow-up questions based on code patterns
• Natural Conversation: Human-like interaction flow with voice integration

3. Code Quality Assessment

• Style Analysis: Adherence to coding standards (PEP 8, ESLint, etc.)
• Readability Scoring: Code clarity and maintainability assessment
• Methodology Detection: Problem-solving approach identification
• Optimization Suggestions: Performance improvement recommendations

4. Comprehensive Reporting

• Visual Dashboard: Clean, intuitive metrics presentation for hiring managers
• Performance Summary: Human-readable candidate evaluation reports
• Session Playback: Complete interview replay with code evolution timeline
• Comparative Analysis: Candidate ranking and skill comparison tools

5. Interactive User Experience

• Intuitive Interface: Seamless coding environment with real-time feedback
• Voice Integration: Audio input/output for natural conversation
• Multi-language Support: Support for Python, JavaScript, Java, C++, and more
• Accessibility: Screen reader compatible and keyboard navigation

🏗️ Architecture

Frontend Stack

• React 18 with TypeScript for type safety
• Vite for fast development and building
• Tailwind CSS for responsive, utility-first styling
• Monaco Editor for advanced code editing with syntax highlighting
• WebRTC for real-time audio communication
• Framer Motion for smooth animations and transitions

Backend Services

• Supabase for real-time database and authentication
• Edge Functions for AI processing and code analysis
• WebSocket connections for real-time communication
• Docker containers for secure code execution sandboxing

AI & Analysis

• OpenAI GPT-4 for intelligent interviewer responses
• Custom AST Parser for code structure analysis
• Static Analysis Tools (ESLint, Pylint) for quality assessment
• Performance Profilers for execution metrics

📊 Evaluation Metrics

Dimension	Weight	Description
Conversational Intelligence	30%	Natural interaction, adaptive responses, context awareness
Analysis & Feedback Quality	25%	Code assessment accuracy, hint relevance, insight depth
Performance & Reliability	20%	System speed, stability, scalability, uptime
User Experience	15%	Interface intuitiveness, accessibility, workflow efficiency
Innovation & Creativity	10%	Novel features, creative solutions, technical excellence

🎯 Key Differentiators

1. Adaptive Intelligence: Dynamic problem difficulty and personalized interview flow
2. Multi-modal Interaction: Seamless text and voice communication
3. Real-time Analysis: Instant code quality and performance feedback
4. Comprehensive Assessment: Beyond correctness to problem-solving methodology
5. Scalable Architecture: Handles hundreds of concurrent interviews
6. Bias Mitigation: Consistent, objective evaluation criteria

🔧 Development Setup

Prerequisites

• Node.js 18+ and npm/yarn
• Supabase account for database services
• OpenAI API key for AI functionality
• RapidAPI account for Judge0 code execution (or self-hosted Judge0)

Installation

# Clone and install dependencies
npm install

# Start development server
npm run dev

# Run type checking
npm run typecheck

# Build for production  
npm run build

Environment Variables

# Supabase Configuration
VITE_SUPABASE_URL=your_supabase_url
VITE_SUPABASE_ANON_KEY=your_supabase_anon_key

# AI Services
VITE_OPENAI_API_KEY=your_openai_api_key

# Judge0 Code Execution (Choose one option)
# Option 1: RapidAPI (Recommended)
VITE_RAPIDAPI_KEY=your_rapidapi_key_here
VITE_JUDGE0_API_URL=http://localhost:3258

# Option 2: Self-hosted Judge0
# VITE_JUDGE0_API_URL=http://localhost:2358
# VITE_JUDGE0_API_KEY=your_api_key_here

📖 Judge0 Setup: For detailed instructions on setting up Judge0 for code execution, see JUDGE0_SETUP.md

📁 Project Structure

src/
├── components/           # Reusable UI components
│   ├── interview/        # Interview-specific components
│   ├── dashboard/        # Admin dashboard components
│   ├── common/          # Shared components
│   └── ui/              # Base UI components
├── services/            # API and external service integrations
│   ├── supabase.ts      # Database client
│   ├── ai-interviewer.ts # AI service integration
│   └── code-analysis.ts  # Code evaluation engine
├── hooks/               # Custom React hooks
├── utils/               # Utility functions and helpers
├── types/               # TypeScript type definitions
└── pages/               # Route components
    ├── Interview.tsx     # Main interview interface
    ├── Dashboard.tsx     # Admin/HR dashboard
    └── Reports.tsx       # Detailed reporting view

🎮 Usage Flow

Quick Demo

1. Launch Application: npm run dev and visit http://localhost:5173
2. Try Judge0 Demo: Click "Launch Demo" to test code execution
3. Test Different Languages: Switch between JavaScript, Python, Java, etc.
4. Experience Features: Real-time execution, error handling, and test validation

For Candidates

1. Join Interview: Enter session code or direct link
2. Voice Setup: Audio/microphone permission and testing
3. Problem Introduction: AI explains the coding challenge
4. Live Coding: Real-time code analysis and feedback with Judge0 execution
5. Interactive Discussion: Voice-based Q&A and hints
6. Session Complete: Final assessment and next steps

For Interviewers/HR

1. Create Session: Set up interview parameters and problems
2. Monitor Live: Real-time candidate performance tracking
3. View Analysis: Comprehensive code quality metrics
4. Generate Report: Automated assessment summary
5. Review Playback: Complete session replay and analysis

🔒 Security & Privacy

• Sandboxed Execution: Isolated code execution environment
• Data Encryption: End-to-end encryption for all communications
• GDPR Compliance: Privacy-first data handling and storage
• Access Controls: Role-based permissions and authentication
• Audit Logging: Complete interview session tracking

📈 Performance Optimization

• Code Splitting: Lazy loading for optimal bundle sizes
• WebSocket Pooling: Efficient real-time connection management
• Caching Strategy: Smart caching for repeated analysis patterns
• CDN Integration: Global content delivery for low latency
• Progressive Enhancement: Graceful degradation for various devices