Multimodal Intelligence for Recruitment Process Automation

This project presents a Multimodal Intelligence for Recruitment Process Automation platform designed to transform traditional hiring into an efficient, unbiased, and intelligent end-to-end process. The system integrates advanced Artificial Intelligence and Machine Learning techniques across five key recruitment stages: semantic resume screening using Sentence-BERT (SBERT), aptitude and cognitive assessment using a Gradient Boosting Classifier (GBC), technical coding skill evaluation using Code2Vecbased embeddings, personalized AI interview question generation using a fine-tuned T5 Transformer, and multimodal behavioral analysis combining MFCC with BiLSTM for voice assessment and a Spatiotemporal Attention Network for facial expression and engagement detection. All modules are unified within a Flask-based web platform that provides real-time dashboards, candidate rankings, and automated evaluation reports for HR teams. Experimental results demonstrate high accuracy in candidate evaluation across all modules, making the system a reliable and scalable solution for modern recruitment automation.

Keywords: Artificial Intelligence, Machine Learning, Natural Language Processing, Sentence-BERT, T5 Transformer, Gradient Boosting Classifier, Code2Vec, BiLSTM, MFCC, Deep Learning, Recruitment Automation, Resume Screening, Voice Analysis, Facial Expression Recognition, HR Avatar.

Aravind M, Baskar G, Hari Babu L, T. Kanchana, Dr. N. Shunmuga Karpagam (2026). Multimodal Intelligence for Recruitment Process Automation. International Journal of Innovative Research in Multidisciplinary Education & Technology (IJIRMET), 11(3).