Placementor

AI-Powered Voice Interview Coach

Comment

🚀 Placementor — AI-Powered Voice Interview Coach

Built Using Murf Falcon

Deployed Link: link

📚 Table of Contents

Inspiration

Real interviews depend on communication skills — yet most prep platforms are entirely text-based. Millions of students and job-seekers struggle not because they lack knowledge, but because they lack a realistic space to practice speaking under pressure.

We wanted to build a platform where you don't just read interview questions — you hear them, answer them out loud, and get instant feedback. That gap between text-based prep and real interview dynamics was our starting point for Placementor.

What it does

Placementor is a voice-first AI interview training platform that simulates a real interview experience end-to-end:

  • 🎤 Hear questions narrated by Murf Falcon TTS in a natural interviewer voice
  • 🗣️ Speak your answers captured live via Automatic Speech Recognition (ASR)
  • 📊 Receive instant structured AI feedback on structure, clarity, relevance, and depth
  • 📄 Upload your resume for personalized, role-specific question generation
  • 🔁 Repeat or proceed based on AI recommendations for mastery learning

It supports company-specific, role-specific, and round-specific interview simulations — making prep genuinely targeted and immersive.

How we built it

Placementor is built on a modular dual-agent architecture with a voice layer on top:

  • Frontend: React.js + Tailwind CSS for a clean, responsive UI
  • Backend: Python + Flask for fast real-time API handling
  • Voice Layer:
    • Murf Falcon TTS API for low-latency, natural question narration
    • ASR for live speech-to-text answer capture
  • AI Agents: Built using the Agno framework + Gemini, split into four dedicated agents:
    • Resume Parser — extracts structured data from uploaded PDFs
    • Interview Planner — generates progressively difficult, personalized question sets
    • Question Fetcher — retrieves questions by serial number for smooth voice flow
    • Feedback Agent — scores answers and provides detailed improvement guidance

The seamless loop of listen → speak → analyze → respond was made possible by Murf Falcon's stable, low-latency API output integrating cleanly with our Flask backend.

Challenges we ran into

  • Synchronizing TTS narration with ASR capture without overlap or lag
  • Ensuring low-latency voice output so the conversation felt natural, not robotic
  • Designing structured AI scoring that was consistent and not arbitrary across sessions
  • Parsing diverse resume formats reliably into clean, structured JSON
  • Balancing question difficulty progression so interviews felt realistic rather than random
  • Building a smooth repeat-or-proceed logic that didn't disrupt the interview flow

Accomplishments that we're proud of

  • ✅ Built a fully voice-first interview simulation — hear the question, speak the answer, get feedback
  • ✅ Achieved seamless Murf Falcon + ASR integration with real-time response
  • ✅ Designed a modular 4-agent architecture that is clean, maintainable, and extensible
  • ✅ Delivered company and role-specific question generation from resume context
  • ✅ Implemented structured AI feedback scoring (1–10) with ideal answer comparisons
  • ✅ Deployed a fully functional live product at place-mentor-murf-ai.vercel.app

What we learned

  • Voice-first UX requires a fundamentally different design mindset than text-based apps — latency is everything
  • Structured agent outputs (clean JSON) are far more reliable than freeform AI responses for downstream processing
  • Murf Falcon's stable API made real-time TTS integration far more practical than expected
  • Resume parsing needs robust fallback handling given the variety of real-world formats
  • Modular agent separation makes the system significantly easier to debug, test, and extend

What's next for Placementor

  • [ ] Multiple voice personalities — choose HR, technical, or panel interviewer tones
  • [ ] Mock HR stress testing — pressure-based question sequences
  • [ ] Multilingual interview modes — practice in regional languages
  • [ ] Company-specific interview packs — curated question banks per company
  • [ ] Behavioral + Technical + System Design rounds — full interview cycle support
  • [ ] Tone and pacing analytics — feedback on how you speak, not just what you say
  • [ ] Progress tracking dashboard — visualize improvement over sessions

About

Placementor is a next-generation voice-first interview training platform designed to help users practice, improve, and perfect their interview performance. It integrates:

✅ Murf Falcon TTS for ultra-natural, real-time question narration ✅ Automatic Speech Recognition (ASR) for live conversational interaction ✅ AI Agents for adaptive question generation & feedback

Features

  • 🎤 Voice-first Interview Simulation using Murf Falcon TTS
  • 🗣️ Real-time ASR-based Response Capture
  • 📄 Resume Upload & Parsing
  • 🧠 AI Interview Planner generates role/company-specific questions
  • ❓ Dynamic Question Fetching
  • 📊 AI Feedback Agent provides structure, clarity & relevance scores
  • 🔁 Repeat Question Logic for mastery learning
  • ⚡ Fast Flask Backend for Real-Time Conversations

Tech Stack

  • Frontend: React.js, Tailwind CSS
  • Backend: Python, Flask
  • Voice Tech:
    • Murf Falcon TTS API
    • ASR (Automatic Speech Recognition)
  • AI Agents Framework: Agno + Gemini
  • Others: Axios, JSON, Git

Agents

Resume Parser

  • Extracts structured resume data: name, email, phone, skills, projects, education, experience.
  • Uses Gemini via Agno Framework.
  • Returns clean JSON with parsed content.

Interview Planner

  • Generates progressively difficult interview questions.
  • Personalized based on company, role, resume, and round.
  • Outputs a detailed JSON plan.

Question Fetcher

  • Retrieves specific questions by serial number.
  • Ensures smooth integration with the voice agent step.

Feedback Agent

  • Evaluates user answers using AI scoring parameters: structure, clarity, relevance, depth.
  • Provides:
    • Score (1–10)
    • Detailed feedback
    • Corrected ideal answer
    • Repeat-or-proceed recommendation

Usage

  1. Upload Resume (PDF)
  2. Select Target Company + Role + Round
  3. Start Voice Interview
    • Murf Falcon narrates questions
    • ASR captures user answers
  4. Get Instant Feedback
  5. Repeat or Proceed Based on AI Recommendations

Developed By

Built With

+ 8 more
Share this project:

Updates