AgentVox

Adaptive Multi-Agent Voice AI Interview Platform

AgentVox is an AI-powered voice interview platform that conducts adaptive, evidence-based interviews through real-time conversation.

Unlike traditional interview systems that rely on fixed questions and surface-level signals, AgentVox adapts its questions based on the User's responses and evaluates candidates using job-relevant evidence.

✨ Core Idea

AgentVox follows an adaptive interview loop:

Claim → Probe → Evidence → Reasoning → Evaluate

The system analyzes the User's resume and target role, creates a competency-focused interview plan, conducts the interview through real-time voice, and dynamically generates follow-up questions when additional evidence is required.

🏗️ Architecture

AgentVox Architecture

🚀 Key Features

• 🎙️ Real-time voice interviews
• 📄 Resume-based interview context
• 💼 Job description / target-role analysis
• 🧠 Adaptive interview questioning
• 🔍 Claim and evidence detection
• 💡 Reasoning analysis
• 📊 Competency-based evaluation
• 🎯 Confidence-aware assessment
• 🔄 Dynamic follow-up questions
• 📝 Evidence-based interview feedback
• 🔎 Semantic retrieval using vector embeddings

🔄 Interview Flow

Resume + Target Role
        ↓
Context Analysis
        ↓
Competency Mapping
        ↓
Interview Planning
        ↓
Real-Time Voice Interview
        ↓
Candidate Response
        ↓
Claim Detection
        ↓
Evidence Extraction
        ↓
Reasoning Analysis
        ↓
Evidence Check
     ↙       ↘
Insufficient  Sufficient
    ↓             ↓
Follow-up      Evaluation
    ↓             ↓
    └──────→ Next Question
                  ↓
            Final Feedback