Behavioural Analysis models for a project

I'm working on a project to create an AI-powered mock interview platform that evaluates both technical skills and behavioral attributes of candidates. One key feature involves analyzing candidates' behavior using audio and video inputs to provide insights into their performance. Specifically, we need models or frameworks that can handle video analysis for attributes like facial expressions, emotions, posture, gestures, and overall movement, as well as audio analysis for tone, pitch, sentiment, clarity, and fluency. We're aiming for real-time analysis, where the data is processed during the interview without storing raw video or audio. If you know of any robust AI models, libraries, or tools that can handle these tasks effectively, please drop your suggestions! Your recommendations would be incredibly helpful for our project's success.