The traditional faculty- recruitment process involves manual checking of candidates’ credentials and is reliant on the collective experiences and gut feelings. Owing to human limitations, misfit candidates might get selected. Nevertheless, despite these shortcomings, data-driven decision-making is not explored in such setups. The current study attempts a data-driven analysis and has the primary objective to build predictive models for research performances. The study uses faculty data of civil and mechanical engineering departments of selected public engineering colleges. Five data mining methods have been tested here for classification exercises, and two models achieved acceptable predictability. Findings obtained in this study has larger implications for the academic-recruitment process and can be researched further with larger samples.