LumiereAI

JobsMatcher automates end-to-end recruitment through AI voice and video interviews with real-time emotion analysis. The system uses multiple specialized agents to conduct professional interviews, analyze responses, grade candidates, and recommend matching jobs—all deployed on Google Cloud Run.Innovation & Creativity: First system Generates personalized questions from CV using LLM, then it will do a real voice and video interview with the candidate combining real-time facial emotion detection with voice questions to detect his motivation, emotion and answers for the questions. This is an Adaptive interview flow based on candidate responses.
it scan the cv, comes up with questions to ask the candidates, conducts the interview and asks follow questions during the interview. it then saves the transcript and grades the results on a normalized scale using hard and soft skills. another agent uses the grading to come up with the perfect job for you. this perfect job description is encoded into an embedding and matches using FAISS to the other jobs embedding. it finally proposes 3 jobs that are the best for you
Afterward it will match it will the best job using FAISS vector search (sub-ms job matching)

Everything was created during the hackathon