• 

  Landing

• 

  Start Interview

• 

  n8n

WiseHire - Hackathon Journey

🌟 Inspiration

The job market is brutal. We've all been there - staring at a job posting, wondering what questions they'll actually ask, practicing with generic interview prep that feels nothing like the real deal. Traditional interview preparation tools are outdated, generic, and frankly... boring! 😴

We dreamed of a world where:

• Real market data drives your prep (not some dusty question bank from 2015)
• AI avatars conduct lifelike interviews that actually feel engaging
• Your resume becomes the foundation for personalized questions
• 30 seconds is all it takes to generate a complete interview session

WiseHire was born from the frustration of countless developers who deserved better than "Tell me about yourself" for the millionth time. We wanted to revolutionize interview preparation by combining cutting-edge AI technology with real-time market intelligence.

🎯 What it does

WiseHire is your AI-powered interview preparation companion that brings together cutting-edge technology to create the most realistic and relevant interview experience possible:

🔍 Smart Question Generation

• Resume Analysis: Upload your PDF resume and watch our AI extract key skills, experiences, and achievements to craft personalized questions using Cerebras AI
• Real-Time Market Intelligence: We scrape live Google job postings using Search API to ensure your questions reflect what companies are actually asking right now
• Dual Input Modes: No resume? No problem! Manually input job descriptions and get equally powerful question generation

🤖 Lifelike AI Interviews

• Voice Interaction: Speak your answers naturally using AssemblyAI for speech-to-text conversion - no more typing responses into boring text boxes
• AI-Powered Text-to-Speech: Natural conversation flow with Murf AI generating realistic voice responses
• Adaptive Conversations: Our AI interviewer follows up on your responses, creating realistic interview dynamics

📊 Intelligent Feedback System

• Performance Analysis: Get detailed breakdowns of your interview performance powered by Cerebras AI
• Personalized Suggestions: Receive actionable improvement recommendations based on your specific responses
• Progress Tracking: Monitor your improvement over time with comprehensive interview history stored in Convex database

🛡️ Enterprise-Grade Features

• Secure Authentication: Clerk-powered user management with personalized dashboards
• Rate Limiting: Arcjet integration ensures fair usage and system stability with credit-based system
• Cloud Storage: ImageKit handles your resume uploads securely and efficiently

🏗️ How we built it

The Tech Stack That Powers the Magic

• Frontend: Next.js 15 with React 19 & TypeScript for blazing-fast performance
• Styling: Tailwind CSS with custom shadcn/ui components for a beautiful, responsive UI
• Database: Convex for real-time data synchronization and seamless user experiences
• Authentication: Clerk for secure, hassle-free user management
• AI Integration: Cerebras AI for question generation and feedback analysis
• Voice Services: Murf AI for text-to-speech, AssemblyAI for speech-to-text
• File Storage: ImageKit for optimized PDF resume handling
• Rate Limiting: Arcjet for API protection and credit management

The n8n Workflow Revolution 🔄

Our secret sauce is the sophisticated n8n automation workflow that orchestrates the entire question generation process:

📥 Webhook Trigger
    ↓
🔍 Resume Detection Logic
    ↓
📋 Branch Processing:
├── 📄 With Resume Path:
│   ├── 📤 Upload to ImageKit
│   ├── 🤖 AI Resume Parsing (Cerebras Llama)
│   ├── 🌐 Google Job Scraping (Search API)
│   └── 💡 Question Generation (Cerebras Llama)
│
└── ✍️ Manual Input Path:
    ├── 🌐 Google Job Scraping (Search API)
    ├── 🧹 Data Cleaning & Optimization
    └── 💡 Question Generation (Cerebras Llama)

Key Workflow Features:

• Cerebras AI Integration: Llama model for both question generation and resume parsing
• Real-Time Data: Search API scrapes fresh Google job postings based on job titles
• Smart Branching: Different processing paths optimize for resume vs. manual input
• Data Optimization: Custom code nodes clean and structure scraped data for maximum relevance

Component Architecture

• CreateInterviewDialog: Tabbed interface for resume upload and job description input
• Dashboard: Personalized interview history with beautiful card layouts
• Interview Interface: Real-time voice processing with AI-powered conversations
• Feedback System: Comprehensive performance analysis and improvement suggestions

🚧 Challenges we ran into

The Voice Integration Complexity 😅

Integrating real-time voice processing was like trying to conduct an orchestra - technically complex but incredibly rewarding! We had to:

• Master AssemblyAI's WebSocket streaming for real-time speech-to-text
• Handle voice activity detection seamlessly
• Ensure natural conversation flow between human and AI
• Debug audio processing across different browsers and devices
• Implement proper error handling for network interruptions

The Job Scraping Puzzle 🧩

Getting consistent, high-quality data from Google job postings required careful engineering:

• Navigate different HTML structures across various job posting formats
• Handle dynamic content loading and JavaScript-heavy pages
• Implement robust error handling for failed scraping attempts
• Parse and clean inconsistent data formats into usable information

The n8n Workflow Complexity 🌊

Building a workflow that handles both resume and manual inputs while maintaining data quality was like conducting an orchestra:

• Designing conditional logic that routes data correctly based on user input
• Ensuring Cerebras AI agents receive properly formatted input for optimal results
• Handling file uploads and processing in a distributed system
• Managing API rate limits across multiple services (Cerebras, Murf, AssemblyAI)
• Implementing proper error handling and retry mechanisms

Real-Time State Management ⚡

Keeping the UI responsive while complex AI processing happens in the background required:

• Implementing progressive loading states and user feedback
• Managing complex state across multiple components during voice interactions
• Handling WebSocket connections for real-time speech processing
• Ensuring graceful error recovery and informative user notifications
• Synchronizing state between voice services and UI components

Credit System Implementation 💳

Implementing a fair usage system with Arcjet rate limiting:

• Designing token bucket algorithms for different user tiers
• Handling rate limit errors gracefully without breaking user experience
• Implementing user feedback for credit consumption and limits

🏆 Accomplishments that we're proud of

🎨 The User Experience Revolution

We didn't just build another interview tool - we crafted an experience. From the moment users land on our page to their final feedback report, every interaction feels polished, intuitive, and engaging. Our responsive design works flawlessly across all devices with beautiful dark/light theme support.

🤖 AI Integration Mastery

Successfully integrating multiple AI services (Cerebras, Murf, AssemblyAI) into a cohesive workflow that feels seamless to users. The fact that users can have natural voice conversations while getting market-relevant questions powered by real-time data is pretty magical! ✨

⚡ Performance Optimization

Despite the complex backend processing, we maintained lightning-fast user interactions:

• Sub-second page loads with Next.js 15 optimization
• Real-time data synchronization with Convex database
• Efficient file handling with ImageKit CDN
• Smart caching strategies for repeated requests
• Optimized voice processing with minimal latency

🔒 Production-Ready Architecture

Built with enterprise-grade security and scalability in mind:

• Comprehensive error handling and user feedback systems
• Rate limiting and abuse prevention with Arcjet
• Secure file uploads and data processing
• Proper authentication and user session management
• WCAG compliant accessibility features

🌐 Real Market Intelligence

The integration with Search API to scrape live Google job postings means our questions are always current and relevant - not some stale database from years ago! This real-time market data makes our interview prep incredibly valuable.

🔄 Sophisticated Automation

Our n8n workflow automation handles complex business logic seamlessly:

• Dual-path processing based on user input type
• Intelligent data cleaning and optimization
• Robust error handling and retry mechanisms
• Scalable architecture that can handle multiple concurrent requests

📚 What we learned

🔄 The Power of Workflow Automation

n8n taught us that complex business logic doesn't have to live in your application code. By offloading the heavy lifting to automated workflows, we kept our Next.js app clean and focused on user experience while the backend handles the complexity. This separation of concerns made our codebase more maintainable and scalable.

🎭 Voice Interactions Change Everything

Working with AssemblyAI and Murf showed us that voice interactions aren't just about functionality - they're about creating genuine, natural experiences. The ability to speak your answers and hear AI responses makes interview prep feel real and engaging rather than mechanical.

📊 Data Quality > Data Quantity

Initially, we tried to scrape everything from job postings. We learned that carefully curated, cleaned data produces much better interview questions than raw, unprocessed information. Quality preprocessing with proper data cleaning is everything!

🚀 Progressive Enhancement is Key

Starting with core functionality and progressively adding features allowed us to ship faster while continuously improving the user experience. Building the MVP first and then enhancing with voice capabilities was the right approach.

🔐 Security from Day One

Implementing proper authentication with Clerk, rate limiting with Arcjet, and data validation from the beginning saved us countless headaches. These enterprise-grade security measures made our hackathon project production-ready.

🤖 AI Service Integration Complexity

Each AI service has its own quirks and requirements. Learning to handle different API patterns, rate limits, and error conditions across Cerebras, Murf, and AssemblyAI taught us valuable lessons about building resilient AI-powered applications.

📱 Responsive Design Matters

With voice interactions, users might be on mobile devices. Ensuring our interface works seamlessly across all screen sizes and input methods was crucial for the user experience.

🚀 What's next for WiseHire

🎯 Enhanced AI Capabilities

• Advanced Feedback Analysis: Deeper performance insights using Cerebras AI's advanced models
• Adaptive Difficulty: AI that adjusts question complexity based on user performance
• Industry-Specific Training: Specialized models for different job sectors and roles

🌍 Multi-Language Support

Expanding beyond English to support global job markets:

• Multi-language voice processing with AssemblyAI and Murf
• Culturally-aware interview scenarios
• Region-specific job market data integration

📱 Mobile App Experience

Native iOS and Android apps for interview prep on-the-go:

• Offline practice mode for questions
• Push notifications for new market trends
• Seamless sync with web platform

🤝 Advanced Collaboration Features

• HR Integration: Custom interview templates for companies
• Peer Learning: Community features for shared learning
• Expert Feedback: Integration with human interview coaches

📈 Advanced Analytics Dashboard

• Market Trend Analysis: Real-time insights into job market demands
• Skill Gap Analysis: Identification of areas for improvement
• Success Tracking: Long-term career progress monitoring
• Competitive Analysis: Benchmarking against industry standards

🔗 Integration Ecosystem

• Job Board Integration: Direct connection with major job platforms
• Calendar Sync: Schedule practice sessions around real interviews
• LinkedIn Integration: Portfolio and experience sync
• ATS Optimization: Resume optimization for specific company systems

🧠 Advanced Voice Features

• Emotion Detection: AI that adapts based on candidate stress levels in voice
• Accent Training: Pronunciation and clarity improvement
• Real-time Coaching: Live suggestions during practice sessions

🔒 Enterprise Features

• Team Dashboards: Organization-wide interview preparation tracking
• Custom Branding: White-label solutions for HR departments
• Advanced Analytics: Detailed reporting and insights for organizations
• API Access: Integration capabilities for third-party platforms

Built With

• assembly
• convex
• murf
• n8n
• nextjs