Banks have started adopting machine learning to reinvent their business functions and improve their decision-making capabilities. However, operationalizing machine learning. i.e. converting machine learning experiments into sustainable production-grade applications, remains a key challenge, which limits banks from realizing the true business value of machine learning. Despite its significance, the scholarly literature on machine learning operationalization is scant and predominantly technical. The main objective of this research is to develop a socio-technical framework, that supports the understanding and implementation of the machine learning operationalization process in the banking sector. This research combines an extensive literature study and 15 expert interviews to identify nine socio-technical factors that influence the operationalization of machine learning applications. The identified factors are then validated and applied to a real-world context through a case study analysis. The findings suggest that risk management is one of the most crucial yet challenging aspects of the process. To investigate this further, the research analyzes the socio-technical challenges of risk management and proposes four strategic guidelines to address the same. This leads to the development of a conceptual model which illustrates how factors such as shared knowledge and controls reduce the challenges of risk management and thereby support the machine learning operationalization process.