Intra AI — Adaptive AI Recruitment and Interview Platform
Link to open source: https://github.com/VIsHnu1762/intra-ai
Intra AI — Real-Time AI Interview Platform
Intra AI is an AI-powered recruitment platform that helps companies screen, interview, and evaluate candidates more efficiently.
The key part of our solution is the real-time voice interview experience built using Agora Conversational AI. Candidates can join an interview and speak naturally with AI interview agents. Agora manages the live audio conversation, speech recognition, turn detection, and text-to-speech, while our FastAPI backend manages the interview intelligence and business logic.
Our Gemini-powered M1 layer analyzes every candidate response, identifies skills and evidence, detects gaps or contradictions, and recommends meaningful follow-up questions. A LangGraph-based orchestrator controls the interview flow, adjusts question difficulty, and can hand off the conversation between technical and product-focused interview agents.
After the interview, recruiters receive a structured report containing scores, strengths, improvement areas, competency coverage, and evidence from the candidate’s responses.
We built Intra AI to reduce repetitive recruitment work while keeping human recruiters in control. It is especially useful for startups, small companies, and recruitment teams that need a scalable, consistent, and interactive hiring process.
Intra AI also includes Morgan, an AI assistant for recruiter operations, and Taylor, a practice coach that helps candidates prepare for interviews and improve their performance.
The demo showcases the complete journey: creating a job, selecting a candidate, conducting an Agora voice interview, adapting the conversation, and generating the final assessment report.
Supporting Links
- Source Code: https://github.com/VIsHnu1762/intra-ai
- Technical Documentation: https://drive.google.com/drive/folders/1r6zGGitiNHcDB2Zor8bMFtMw_7waHy3p?usp=drive_link
- Demo Video: https://www.youtube.com/watch?v=2Wvalvci71E
This build was uploaded as a hackathon project






