HireWise

AI Interviews. Real Market Data. Unstoppable Preparation.

Comment

HireWise - AI Interviews. Real Market Data. Unstoppable Preparation.

Inspiration

The job market is more competitive than ever, yet most interview preparation tools rely on outdated, generic questions that don't reflect what employers are actually asking today. We were inspired to create a solution that combines the power of real-time job market intelligence with AI technology to give candidates a genuine competitive edge. The idea was born from experiencing firsthand how traditional interview prep fails to capture the dynamic nature of today's hiring landscape.

What it does

HireWise is an AI-powered interview preparation platform that revolutionizes how candidates practice for job interviews:

  • Real-Time Market Intelligence: Scrapes current LinkedIn job postings to generate interview questions that reflect actual market demands
  • Lifelike AI Interviews: Uses Akool's streaming avatar SDK to conduct realistic practice sessions with voice interaction
  • Adaptive Question Generation: Creates personalized questions based on uploaded resumes or manual job descriptions
  • Intelligent Feedback System: Provides comprehensive performance analysis and improvement suggestions
  • Secure User Management: Implements Clerk authentication with personalized interview history tracking

The platform ensures candidates practice with questions employers are asking today, not generic questions from outdated databases.

How we built it

Technology Stack

  • Frontend: Next.js 15 with React 19, TypeScript, and Tailwind CSS
  • Backend: Next.js API routes with Convex database
  • AI Integration: Akool Streaming Avatar SDK for lifelike interviews
  • Authentication: Clerk for secure user management
  • File Storage: ImageKit for PDF resume uploads
  • Rate Limiting: Arcjet for API protection and credit management

n8n Workflow Architecture

We built sophisticated n8n workflows that demonstrate the "unstoppable workflow" concept:

  1. Intelligent Branching Logic:

    • Detects whether user uploaded a resume or provided manual job details
    • Routes to appropriate processing path

  2. Bright Data Integration:

    • Node: "Access and extract data from a specific URL" (configured with web_unlocker1 zone)
    • Purpose: Scrapes LinkedIn job postings based on job titles
    • Data Processing: Custom code nodes clean HTML and optimize scraped content

  3. Multi-AI Processing:

    • Google Gemini: Parses and extracts resume content
    • OpenAI GPT: Generates contextual interview questions based on market data
    • AI Transform: Processes and structures scraped job market data

  4. Dynamic Data Flow:

    Webhook → Resume Check → Branch Logic:
├── With Resume: Upload → Extract PDF → Parse with Gemini → Scrape LinkedIn → Generate Questions
└── Without Resume: Scrape LinkedIn → Clean Data → Generate Questions

Key Implementation Features

  • Resume Processing: PDF upload, extraction, and AI-powered content analysis
  • Voice Interaction: Real-time conversation with AI avatars using Akool's streaming SDK
  • Credit System: Arcjet-based rate limiting with user credit management
  • Responsive Design: Mobile-first UI with dark/light theme support
  • Interview Management: Complete CRUD operations for interview sessions and history

Challenges we ran into

  1. Complex Workflow Orchestration: Building branching logic in n8n that could handle both resume-based and manual job description flows required careful planning and testing.

  2. Real-Time Data Integration: Ensuring Bright Data's scraped LinkedIn job postings were properly cleaned, formatted, and integrated into the AI question generation process.

  3. AI Avatar Synchronization: Implementing smooth voice interactions with Akool's streaming avatars while maintaining conversation context and natural flow.

  4. Rate Limiting Strategy: Balancing user experience with API cost management using Arcjet's credit system without frustrating legitimate users.

  5. Dynamic Question Generation: Making AI-generated questions feel natural and relevant based on both scraped market data and user-specific information (resume/job description).

Accomplishments that we're proud of

  1. Seamless Integration: Successfully integrated n8n AI Agent nodes with Bright Data's verified node to create a truly intelligent workflow that enhances AI capabilities with real-time web data.

  2. Production-Ready Architecture: Built a scalable Next.js application with proper error handling, authentication, database management, and API rate limiting.

  3. Innovative Use of Market Data: Created a system where interview questions stay current with actual job market trends, giving users a real competitive advantage.

  4. Voice-First Experience: Implemented natural voice conversations with AI avatars, making practice sessions feel like real interviews.

  5. Comprehensive User Journey: From resume upload to interview completion to feedback analysis - every step is thoughtfully designed and implemented.

  6. Technical Excellence: Successfully orchestrated multiple APIs (n8n, Bright Data, OpenAI, Akool, ImageKit, Clerk) into a cohesive, reliable system.

What we learned

  1. n8n's Power: Discovered how n8n's workflow automation can handle complex branching logic and multi-step processes, making it perfect for orchestrating AI and web scraping operations.

  2. Real-Time Data Value: Learned that combining real-time web data with AI dramatically improves the relevance and effectiveness of generated content.

  3. User Experience Design: Understood the importance of progressive disclosure and adaptive UI - showing different interfaces based on user choices (resume vs. manual input).

  4. API Orchestration: Gained expertise in coordinating multiple external services while maintaining performance and reliability.

  5. Market Research Integration: Discovered how web scraping can provide context that makes AI responses significantly more valuable and current.

What's next for HireWise

Immediate Enhancements

  1. Extended Market Coverage: Expand Bright Data scraping to include Indeed, Glassdoor, and company career pages for broader market intelligence
  2. Industry Specialization: Create specialized interview modules for different sectors (tech, finance, healthcare, etc.)
  3. Performance Analytics: Add detailed analytics dashboard showing improvement trends and skill gap analysis

Advanced Features

  1. Company-Specific Preparation: Scrape specific company information to create tailored interview experiences
  2. Salary Negotiation Coaching: Integrate real-time salary data to help candidates prepare for compensation discussions
  3. Multi-Language Support: Expand to support interviews in different languages for global job markets

AI Enhancements

  1. Behavioral Pattern Analysis: Use AI to identify speech patterns, confidence levels, and body language cues
  2. Custom Avatar Training: Allow users to practice with avatars that mimic specific interviewer types
  3. Skill Assessment Integration: Add technical coding challenges and soft skill evaluations

Platform Scaling

  1. Enterprise Solutions: Develop tools for companies to use HireWise for candidate screening
  2. Integration Partnerships: Connect with job boards, ATS systems, and career services
  3. Mobile Application: Create native mobile apps for on-the-go interview preparation

HireWise represents the future of interview preparation - where real-time market intelligence meets cutting-edge AI technology to create truly effective, relevant, and engaging practice experiences.

Technical Architecture

n8n Workflow Details

  • Workflow Name: "BrightData Flow For Interview Generation"
  • Bright Data Nodes: 2 instances configured for LinkedIn job scraping
  • AI Nodes: OpenAI (question generation) + Google Gemini (resume parsing)
  • Processing Nodes: Custom code for HTML cleaning and data optimization
  • Webhook Integration: Seamless connection with Next.js API endpoints

API Endpoints

  • /api/generate-interview-question: Triggers n8n workflow with resume/job data
  • /api/interview-feedback: Processes conversation for AI feedback generation
  • /api/akool-session: Manages AI avatar session creation and management

Database Schema (Convex)

  • UserTable: User profiles and authentication data
  • InterviewSessionTable: Interview records with questions, answers, and feedback
  • Relationships: User-to-interviews with status tracking and history management

Built With

Share this project:

Updates