ML-Driven HR System: Candidate Assessment for Enhanced Recruitment Outcomes

This study developed and implemented a machine learning-driven HR system designed to enhance recruitment outcomes, particularly in online interviews. Challenges such as candidate impersonation, interview question memorization, and deepfake manipulations were addressed. The system incorporated several critical processes utilizing advanced algorithms: pre-interview shortlisting employed Support Vector Machines (SVM) achieving 92% accuracy; cognitive skills assessment utilized adaptive testing methods powered by fine-tuned Large Language Models (LLMs) with 88% accuracy; emotional intelligence evaluation was conducted using deep learning-based emotion detection algorithms attaining 85% accuracy; face verification leveraged Convolutional Neural Networks (CNNs) for identity and deepfake detection with 90% accuracy; online video interviews were analyzed through integrated audio-visual processing techniques; language proficiency testing employed Natural Language Processing (NLP) models reaching 87% accuracy; analysis of thinking patterns was performed using sentiment analysis algorithms with 83% accuracy; and candidate grading was facilitated by ensemble learning methods ensuring consistent and reliable evaluations. The results demonstrated the system’s effectiveness in verifying candidate authenticity and assessing competence, providing a robust solution for modern recruitment challenges.