30. 🎥 Interview Sentiment

Overview

A tool that analyzes recorded video interviews to gauge candidate sentiment, speaking confidence, and emotional cues during responses.

Primary Use Cases

• Talent acquisition assessing candidate demeanor
• Coaching services providing feedback on interview performance
• L&D teams tracking improvement over mock interviews

Key Features

• Video upload or capture via webcam
• ASR transcription: openai/whisper-small for dialogue text
• Sentiment analysis: sentence-level sentiment scoring
• Paralinguistic features: speech rate, filler-word detection
• Visualization: timeline chart of sentiment over session

Tech Stack

• Frontend: React + TypeScript + Tailwind for recording and charts
• Backend: FastAPI (Python)
• AI Models:
  • Sentiment: nlptown/bert-base-multilingual-uncased-sentiment
  • Audio features: Librosa for prosody extraction
• Database: PostgreSQL for storing sessions and scores

Architecture

1. Recording Service: capture video → extract audio track