The interview prep tool that actually understands what you're trying to learn
In today's brutal job market where every ML interview feels like a PhD defense, IntrVu is the intelligent companion that doesn't just throw random questions at you—it learns your weaknesses, adapts to your skill level, and actually helps you improve. Think of it as having a senior ML engineer as your personal tutor, but without the $200/hour price tag.
The modern, intuitive landing page showcasing IntrVu's multi-agent AI interview coaching platform
Most interview prep platforms are static and disconnected from real-world scenarios. They don't adapt to your learning style or provide meaningful feedback.
- Intelligent Question Generation: Creates scenarios based on real interview patterns and your skill level
- Adaptive Learning: Adjusts difficulty based on your performance
- Real-time Feedback: Provides detailed explanations, not just correct/incorrect answers
- Comprehensive Coverage: Covers software development, data science, machine learning, and system design
- Multi-Agent System: Orchestrated AI agents using LangGraph for intelligent question generation
- Production-Ready Performance: Sub-2 second response times with error handling
- Scalable Design: Built to handle thousands of concurrent users
- Real-time Analytics: Track progress with detailed insights
- Modern React Stack: Built with React 18, Vite, and cutting-edge web technologies
- Exceptional UX: Smooth animations, responsive design, and intuitive navigation
- Performance Optimized: Lazy loading, code splitting, and efficient state management
The tech job market is more competitive than ever. Companies want engineers who can think critically, solve complex problems, and communicate effectively across software development, data science, and machine learning domains.
Comprehensive progress tracking with detailed analytics and performance insights
Real-time performance monitoring across different challenge types and time periods
- Critical Thinking Development: Scenario-based challenges across multiple domains
- Communication Skills: Practice explaining complex concepts clearly
- Real-world Problem Solving: Tackle problems that mirror actual industry challenges
- Confidence Building: Regular practice with adaptive difficulty
Interactive scenario-based challenges with real-world problem-solving exercises
- Node.js 18+ and Python 3.9+
- API keys for OpenAI, Anthropic, and Clerk (see setup guides below)
# Clone the repository
git clone <repository-url>
cd InterviewPrepper
# Backend setup
cd backend
python -m venv .venv
source .venv/bin/activate # On Windows: .venv\Scripts\activate
uv pip install -r requirements.txt
# Frontend setup
cd ../frontend
npm install# Terminal 1 - Backend
cd backend
uv run server.py
# Terminal 2 - Frontend
cd frontend
npm run dev- Backend: See backend/README.md for detailed setup and API key requirements
- Frontend: See frontend/README.md for frontend configuration
- Personalized Questions: AI generates questions based on your skill level
- Topic Coverage: Comprehensive coverage of software development, data science, ML, and system design
- Difficulty Adaptation: Automatically adjusts based on your performance
- Instant Feedback: Get detailed explanations of why your answer was right or wrong
- Learning Insights: Understand your strengths and areas for improvement
- Progress Tracking: Monitor your improvement over time
- Smooth Interface: Modern, responsive design that works on any device
- Dark/Light Mode: Comfortable viewing in any lighting condition
- Progress History: Complete record of your practice sessions
- Python: FastAPI for high-performance API development
- LangGraph: Multi-agent orchestration for intelligent workflows
- OpenAI/Anthropic: State-of-the-art AI for question generation and evaluation
- SQLite: Lightweight, reliable database with proper indexing
The system uses a sophisticated multi-agent architecture orchestrated by LangGraph with three specialized workflows:
- Intelligent Question Creation: Generates contextually relevant multiple-choice questions
- Difficulty Calibration: Adapts question complexity based on user skill level
- Domain Coverage: Covers software development, data science, ML, and system design
- Quality Assurance: Ensures all options are plausible with detailed explanations
- Real-world Context: Creates realistic interview scenarios with company/role context
- Progressive Questioning: Builds logical question sequences that test analytical thinking
- Business Constraints: Incorporates real-world limitations (budget, timeline, compliance)
- Comprehensive Evaluation: Provides detailed scoring rubrics and model answers
- Multi-dimensional Scoring: Evaluates technical accuracy (40%), problem-solving (30%), communication (10%), and business impact (20%)
- Detailed Feedback: Provides specific, actionable feedback with improvement suggestions
- Model Answer Generation: Creates polished reference answers for learning
- Realistic Assessment: Mimics actual interview feedback from senior engineers
- LangGraph Orchestration: Manages complex multi-step workflows with error handling
- Consistent Output Format: Ensures structured JSON responses for database storage
- Performance Optimization: Uses cost-effective models with conservative temperature settings
- Scalable Architecture: Easy to extend with additional agents and workflows
- React 18: Latest React with modern hooks and concurrent features
- Vite: Lightning-fast development and build tooling
- Vanta.js: Interactive 3D backgrounds for engaging user experience
- Clerk: Modern authentication with social login support
- Response Time: < 2 seconds for challenge generation
- Concurrent Users: Supports 1000+ simultaneous sessions
- Uptime: 99.9% availability with proper error handling
- Scalability: Easy migration path to cloud infrastructure
- System Design: Multi-agent architecture with proper separation of concerns
- Performance Optimization: Sub-second response times with efficient resource usage
- Error Handling: Graceful degradation and comprehensive error recovery
- Security: JWT validation, rate limiting, and input sanitization
- LLM Integration: Sophisticated use of multiple AI models for different tasks
- Workflow Orchestration: Complex multi-step processes with LangGraph
- Prompt Engineering: Carefully crafted prompts for consistent output
- Evaluation Systems: Intelligent assessment of user responses
- Modern Frontend: React with modern patterns and performance optimization
- RESTful APIs: Clean, well-documented API design
- Database Design: Proper schema design with indexing and optimization
This project demonstrates enterprise-level development practices. Contributions are welcome, but please ensure:
- Code follows established patterns and conventions
- New features include proper error handling
- Performance impact is considered
- Documentation is updated accordingly
This project is licensed under the MIT License - see the LICENSE file for details.
Built with ❤️ by Hetav for the tech community. Because interviews shouldn't be a guessing game!