● At Genarion, I developed a multi-agent interview system that simulates structured job interviews using AI. The system includes multiple agents working together to manage the interview flow. ● The Question Generator Agent creates dynamic interview questions based on the candidate’s profile and previous responses. The Interview Talker Agent conducts the interview, ensuring a natural conversation. A Director Agent monitors the dialogue to keep it relevant, intervening if off-topic discussions arise. ● To enhance realism, I integrated speech-to-text (Whisper API) for candidate responses and text-to-speech (F5 TTS) for the interviewer’s voice. Additionally, candidate responses are translated into English using DeepL for evaluation. The system is powered by Crew AI, Llama 3, and DSPy, enabling intelligent decision-making throughout the process.