Interprepper

An AI-powered interview prepping tool.

Comment

Here’s a complete README written in the requested structure and formatting:

🎯 Interprepper

A FastAPI-powered web application for practicing job interviews with realistic voice interactions powered by AI.

Inspiration

Preparing for interviews can be stressful and time-consuming — especially when balancing realistic practice with constructive feedback. Traditional mock interviews often lack flexibility, personalization, and the ability to simulate real human conversation. Interprepper was built to bridge this gap — creating an intelligent, voice-interactive interview simulator that feels natural, gives tailored feedback, and helps users gain confidence before stepping into the real world.

What it does

Interprepper allows users to practice job interviews through two immersive modes:

  • 🎤 Standard Interview Mode: Users answer AI-generated questions one at a time with realistic voice interactions.
  • 🗣️ Live Interview Mode: A dynamic conversational AI that listens, responds, and asks follow-up questions in real time.

It provides AI-driven feedback, including detailed scoring, performance breakdowns, and improvement suggestions. Users can replay questions, review transcripts, and track progress — making every interview session a meaningful learning experience.

How I built it

The system is built using FastAPI for the backend and TailwindCSS for a responsive, modern frontend.

  • AI Question Generation: Uses Groq’s API to create relevant and varied interview questions tailored to user-selected roles or industries.

  • Voice Interaction:

    • Text-to-Speech (TTS) powered by ElevenLabs reads each question naturally.
    • Speech-to-Text (STT) transcribes user responses in real time.

  • Interview Flow Management: FastAPI handles session management and question sequencing.

  • Audio Processing: Enhances voice clarity through isolation and transcription pipelines.

  • AI Feedback System: Groq analyzes each response for content, delivery, and structure, generating numeric scores and detailed commentary.

Everything is orchestrated in a clean multi-page flow with endpoints for interviews, live sessions, and feedback review.

Challenges I ran into

  • 🧩 Managing conversational context in live mode so the AI could respond naturally without losing track of previous responses.
  • 🎙️ Synchronizing TTS, STT, and FastAPI sessions to ensure seamless transitions between speaking and listening.
  • ⚙️ Optimizing latency — ensuring fast response times for real-time conversation.
  • 🎧 Audio processing complexity, including managing browser MediaRecorder APIs, file uploads, and background noise filtering.
  • 🔐 API key management and ensuring safe local setup while integrating multiple third-party services.

Accomplishments that I'm proud of

  • Built a fully voice-interactive AI interviewer that feels human and conversational.
  • Integrated Groq and ElevenLabs APIs effectively to handle generation, speech, and transcription.
  • Designed a modern, user-friendly interface using TailwindCSS with clear flow and visual feedback for recordings.
  • Implemented live feedback and scoring that feels meaningful and educational.
  • Created a scalable architecture that can easily integrate more AI models or features in the future.

What I learned

  • How to orchestrate multiple AI services (Groq + ElevenLabs) in real-time for a seamless voice-based experience.
  • The intricacies of FastAPI session handling, WebAudio APIs, and voice data processing in browsers.
  • The importance of latency optimization and state management in real-time systems.
  • How conversational context can dramatically change the user experience — especially in simulated interviews.
  • Designing for user psychology — understanding what makes practice feel “real” and effective.

What's next for Interprepper

  • 🚀 Personalized Feedback Models: Fine-tuned AI feedback tailored to specific industries (tech, finance, marketing, etc.)
  • 🌐 Cloud Deployment: Public hosting with user accounts, saved sessions, and analytics dashboards.
  • 🎯 Adaptive Interview Difficulty: AI that adjusts tone and complexity based on user skill level.
  • 🧠 Coaching Insights: Long-term performance tracking and personalized improvement plans.
  • 🤝 Team and Peer Modes: Enable collaborative interview simulations or recruiter-driven sessions.

Built With

Share this project:

Updates