prepwise-ai-interviews

An AI-powered real-time voice interview platform that automates candidate screening, provides instant insights, and helps recruiters scale hiring effortlessly.

Comment

Inspiration

Hiring processes are often time-consuming and repetitive. Recruiters spend countless hours conducting initial screenings, while candidates face long waiting times for feedback. We wanted to build an AI-powered solution that automates first-round interviews and provides real-time, unbiased evaluation.

What it does

Our platform is a real-time AI voice interview agent that:

  • Conducts voice-based interviews with candidates.
  • Asks customized questions based on job roles.
  • Transcribes and analyzes answers in real-time.
  • Generates instant insights for recruiters, including sentiment, confidence, and technical accuracy.
  • Scales hiring by allowing simultaneous candidate interviews without human intervention.

How we built it

  • Frontend: React + TailwindCSS for a smooth, responsive UI.
  • Backend: Node.js/Express with WebSocket support for real-time interactions.
  • AI Models: OpenAI for language understanding, Whisper for speech-to-text, and GPT-based evaluation metrics.
  • Voice Layer: WebRTC + Twilio/Agora for live audio streaming.
  • Database: MongoDB for storing candidate responses and recruiter insights.
  • Deployment: Deployed on Vercel (frontend) and AWS/GCP (backend + AI services).

Challenges we ran into

  • Handling real-time low-latency voice streaming between candidate and AI.
  • Designing unbiased evaluation metrics that are consistent across candidates.
  • Managing scalability when multiple interviews happen in parallel.
  • Integrating multiple APIs smoothly within limited hackathon time.

Accomplishments that we're proud of

  • Built a fully functional end-to-end AI interview platform within hackathon time.
  • Successfully deployed a real-time voice interaction agent.
  • Achieved low latency streaming and instant AI-driven insights.
  • Created a tool that recruiters can genuinely use to save hours of manual effort.

What we learned

  • How to optimize WebRTC and WebSocket for real-time communication.
  • Designing AI pipelines that balance speed, accuracy, and fairness.
  • The importance of UX in making AI tools human-friendly and transparent.
  • How collaboration across frontend, backend, and AI teams accelerates results.

What's next for prepwise-ai-interviews

  • Adding multilingual support so candidates can interview in any language.
  • Expanding to video-based interviews with facial sentiment analysis.
  • Building recruiter dashboards with data-driven insights and analytics.
  • Offering API access for companies to plug the AI interviewer into existing ATS platforms.

Built With

Share this project:

Updates