Machine Learning-Driven Recruitment Recommendation System for Employment in Nigerian Universities

• Murtala ISMAIL

  Author

• Mohammed S. ISMAIL

  Author

• Eli A. JIYA

  Author

This paper presents a Machine Learning-Driven Recruitment Recommendation System tailored to the employment needs of Nigerian universities. Conventional recruitment practices in these institutions remain predominantly manual, characterized by protracted document reviews that are inherently slow, inconsistent, and susceptible to bias. To address these deficiencies, this study proposes an intelligent recruitment recommendation system built upon a Dual-Tower Convolutional Neural Network (CNN) architecture, benchmarked against classical supervised learning models, including Logistic Regression, Support Vector Machines, and Random Forest. The framework methodically matches applicant qualifications with institutional job requirements using a structured synthetic dataset covering academic credentials, fields of specialization, class of degree, relevant experience, and research output. The proposed Dual-Tower CNN achieved outstanding classification performance, attaining 99.1% validation accuracy, an Area Under the ROC Curve (AUC) of 0.9997, and an Average Precision (AP) of 0.9999. The system incorporates a score stabilization mechanism to ensure meaningful visual output for end users through a collaborative recommendation interface. The results confirm that deep learning-based architectures offer an accessible, transparent, and unbiased mechanism for restructuring academic recruitment in Nigeria.

• FJET_21_46_46

Copyright (c) 2026 FUDMA Journal of Engineering and Technology

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.