Machine learning algorithms can be utilized to evaluate the performance of employees in an organization. In particular, supervised algorithms such as classifiers can be deployed to handle the vital task of predicting employee ratings based on historical data. In this paper, an ensemble of stacked classifiers are used as the base models and gradient boosted classifier as the Meta—model, combined to predict their performance ratings that are optimized using optimization framework—Optuna. Optuna based on Bayesian model, automates hyper parameter tuning process. Finally, to understand how each of the base models contribute to the prediction that is forwarded to the meta—model, explainable AI feature SHAP (Shapely Additive eXplanations) is incorporated, which helps to visualize the impact of each models in predicting process. The proposed model is subjected to various metrics computations which emphasizes the efficacy of the model. The findings have implications for human resource management, enabling organizations to make informed data—driven decisions about employee performance improvement and development.