基于多模态大模型的智能面试系统设计与实现OA

Aiming at the problems of high labor costs,strong subjectivity in evaluation during enterprise recruitment,and job seekers'lack of immersive simulation training platforms,this paper designs and implements an intelligent interview system based on multimodal Large Language Model.Adopting a dual-backend architecture featuring decoupling between Java business middle platform and Python microservices,the system integrates large models such as Qwen and Gemini,and constructs three core functional modules:a practical programming sandbox,multi-role voice interview,and multimodal comprehensive evaluation.Test results show that the system runs stably:the programming sandbox can execute code safely and conduct code quality analysis;the voice interview module can simulate conversations in various workplace styles;finally,the system can generate structured evaluation reports integrating multi-dimensional information.This system provides a feasible solution for full-process intelligent talent assessment and training,and has reference value for promoting the engineering application of AI-enabled interviews.