The proliferation of online job websites has eased the difficulties in hiring and applying for jobs globally. Unfortunately, the risk of defrauding desperate job seekers exists with malicious recruiters taking advantage of the loopholes in the online recruitment process. The reactive approach to detecting online job fraud and the subsequent warnings on reputable job websites hasn't curtailed this spiteful act. The purpose of the study is to propose a machine learning model for proactive job fraud detection. In building the predictive model, a job fraud dataset from a job advertisement firm in Ghana was utilised. Using the 10-fold and the 5-fold cross-validation techniques, a job fraud detection model was built by comparing conventional and ensemble machine learning algorithms. The machine learning metrics, including accuracy, F1-score and the area under the curve (AUC) value, were reported and discussed. The findings show that the Random Forest traditional algorithm, with an accuracy of 91.86%, is best suited for the dataset. The investigation further indicates that information gain and chi-square feature selection mechanisms decreased classification accuracy marginally to 91.51%.