This research investigates the application of machine learning classification models for predicting employee job satisfaction, considering demographic, professional, and psychosocial aspects. With a secondary dataset obtained from Kaggle, five supervised learning techniques were applied: Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, and Gradient Boosting. The best model was considered to be Gradient Boosting as it achieved the highest accuracy and F1 score. The model's explainability was enhanced with LIME. LIME enhanced the model's explainability. Stress, work-life balance, and job tenure were identified as the primary factors of job satisfaction. These results support the Job Demands-Resources (JD-R) theory and highlight the model's effectiveness in the hands of HR professionals. The study highlights the need to achieve a balance between predictive accuracy and explainability to ethically align the use of AI in HR analytics, aiming to enhance the well-being of employees and the effectiveness of organizations.