The situation
First-round interviews take recruiter time and vary from interviewer to interviewer.
Interview agent using LLMs
An interview agent that reads a résumé, generates role-specific questions and runs technical and soft-skill interviews by voice or text.
Recruitment and HR
01 · Context
02 · Challenge
What had to be true
Generate relevant questions from each résumé, hold a natural conversation, and score answers consistently for HR to review.
03 · What we built
The system
- LLMs (Llama and Gemini) extract skills, experience and achievements from résumés
- Tailored technical and behavioural questions for each role
- An interactive interview flow that stores real-time feedback and scores
- Soft skills scored on communication, problem-solving and teamwork; technical answers assessed by the model
- Speech-to-text and text-to-speech with OpenAI Whisper for natural conversation
- MongoDB stores résumé data, question sets and interview logs; Python and Flask power the API
04 Architecture
How it works, step by step.
Step through the system, or let it play.
- A résumé is uploaded
- Skills and experience extracted
- Questions generated for the role
- Interview by voice or chat
- Technical and soft skills scored
- Report saved for HR
05 · Outcome
HR and managers get consistent first-round interviews with scores and full logs ready to review.
06 · Stack
Next case · Healthcare · NLPClinical decision support platform