Back to portfolio
Hireflow logo

Hireflow

Voice-based AI Agent for live interviews

One workspace for AI interviews, resume intelligence, and job discovery — a real-time voice AI interviewer that adapts as the conversation happens.

Role Personal projectDuration Month YYYY — Month YYYY
LiveGitHubGitHub
Hireflow landing page

Hireflow runs live, voice-based interviews with an AI interviewer that adapts its questions in real time, transcribes as the conversation happens, and evaluates the candidate asynchronously afterward — aiming to replace static, scripted mock interviews with something that actually reacts to what you say.

Voice pipeline

Built the real-time voice interview pipeline using LiveKit WebRTC for audio transport and Azure OpenAI for speech-to-text, the interview LLM, and text-to-speech, so the interview happens as a live spoken conversation rather than typed Q&A.

Adaptive interview logic

Implemented adaptive questioning so follow-up questions respond to the candidate's previous answers instead of following a fixed script.

Live transcription

Streamed live transcription of the conversation as it happens.

Async evaluation & recording

Built background processing with BullMQ and Redis to handle post-interview evaluation and recording pipelines asynchronously, keeping the live call responsive.

Stack & Learnings

<Real-time/>

  • LiveKit
  • WebRTC

<AI/>

  • Azure OpenAI (STT, LLM, TTS)
  • LangChain

<Backend/>

  • Python
  • Express
  • BullMQ
  • Redis

<Database/>

  • PostgreSQL

<Learnings/>

Coming soon

A Closer Look

Screenshot coming soon

Live interview screen

Screenshot coming soon

Adaptive question flow

Screenshot coming soon

Transcription view

Screenshot coming soon

Evaluation report

Screenshot coming soon

Job discovery / dashboard

Back to portfolio