PrepWise: AI Interview Preparation Platform

Inspiration

The journey to landing a tech job involves countless interviews, but candidates rarely get enough realistic practice. We noticed that most people practice with static question lists or rely on friends who may not provide the structured feedback needed to improve. This gap inspired us to create PrepWise—an AI-powered interview simulator that offers realistic, company-specific interview practice on demand.

What it does

PrepWise provides personalized AI interview simulations for tech roles at various companies:

  1. Customizable Interview Experience: Choose from preset interviews at top tech companies like Google, Meta, and Amazon, or create custom interviews for specific roles.

  2. Interactive Voice Conversations: Engage in realistic spoken interviews with our AI interviewer that adapts to your responses.

  3. Real-time Visual Feedback: See visual indicators when the AI is speaking, with animated avatars that simulate a video interview experience.

  4. Secure Recording & Analysis: All interviews are securely recorded with quantum-inspired encryption for post-interview feedback and analysis.

  5. Personalized Improvement Plan: Receive detailed feedback on your performance with specific improvement suggestions.

How we built it

We developed PrepWise using a modern tech stack:

  • Frontend: Next.js, TypeScript, and Tailwind CSS for a responsive, accessible UI
  • Voice AI: VAPI (Voice API) for natural conversational interviews
  • Analysis Engine: Google's Gemini API for technical response evaluation
  • Security: Custom encryption service for secure storage of interviews
  • Database: MongoDB for user data and interview history
  • Authentication: Firebase for secure user authentication

The core interview experience runs in our Agent.tsx component, which manages the voice interaction, real-time speech detection, and interview flow. We implemented a sophisticated company logo matching system to enhance the visual experience when users select specific companies.

Challenges we ran into

  1. Speech Detection Accuracy: Creating a reliable system to detect when users are speaking required custom audio analysis using Web Audio API.

  2. Consistent Interview Experiences: Ensuring the AI maintained context throughout the interview and provided appropriate questions for each role and level.

  3. Cross-browser Audio Compatibility: Audio capture and processing works differently across browsers, requiring extensive testing and fallbacks.

  4. Secure Storage: Implementing end-to-end encryption for interview recordings while maintaining performance.

  5. UI Responsiveness: Creating animated speaking indicators that felt natural and responsive to actual speech patterns.

Accomplishments that we're proud of

  1. Our real-time speech detection system that shows visual indicators when either the AI or user is speaking.

  2. The seamless integration between VAPI for conversation and Gemini for analysis.

  3. The elegant card-based UI that presents company-specific interviews with appropriate branding.

  4. Implementing a sophisticated company logo matching algorithm that handles various company name formats.

  5. Creating a secure, privacy-focused system that encrypts all interview data.

What we learned

  1. The importance of real-time feedback in creating engaging AI interactions.

  2. How to effectively combine multiple AI models for different aspects of a complex application.

  3. Techniques for processing audio streams in the browser for speech detection.

  4. The challenges of prompt engineering to ensure consistent AI interview behaviors.

  5. How to create animations that enhance the feeling of a natural conversation.

What's next for PrepWise

  1. Expanded Company Coverage: Adding more company-specific interview simulations based on real interview data.

  2. Video Interview Mode: Introducing optional video recording to analyze body language and presentation.

  3. Industry-Specific Tracks: Expanding beyond tech to other industries like finance, healthcare, and marketing.

  4. Interview Marketplace: Allowing interview coaches to create and sell custom interview templates.

  5. Team Assessment: Enabling multi-participant interviews to evaluate teamwork and collaboration skills.

  6. Mobile App: Developing native mobile applications for on-the-go interview practice.

By continuing to refine our AI models and expand our question database, we aim to make PrepWise the gold standard for interview preparation across all industries.

Built With

Share this project:

Updates