Published Paper Details:

Virtual interview systems are increasingly used in recruitment processes, yet they face major challenges in capturing behavioral cues such as sentiment, nervousness, and body language that are essential for fair and accurate candidate evaluations. Current platforms primarily focus on candidate preparation while offering minimal support to interviewers, leading to bias, inconsistency, and overlooked talent. This paper presents DualPrep, an AI powered interview platform designed to empower both candidates and interviewers through real time behavioral analysis and adaptive questioning. The system integrates sentiment analysis via DeepFace, video conferencing through DigitalSamba, natural language generation using Cohere API, and Firebase for data management. Unlike traditional platforms that automate the entire hiring process and often result in trust issues, DualPrep assists interviewers by providing evidence based insights and dynamically generated follow up questions while maintaining human oversight in final decisions. The platform addresses key limitations of existing systems including lack of adaptive questioning, insufficient interviewer support, and poor integration of multimodal analysis. Future enhancements include implementing cheating prevention mechanisms and conducting in depth technical analysis. This research contributes to the growing body of work on AI assisted recruitment by proposing a balanced approach that combines automation with human judgment.