The Modulai team was hired to develop a machine learning model for predicting the final outcome of a transaction. The project included a full specification of the decision engine infrastructure as well as the deployment plan.
The company’s current risk and decision systems were thoroughly researched and new ML-based approach to estimating risk was developed and tested. The central objectives included reducing time to market but without compromising on the upside.
This client is a large fintech in the payments space providing an alternative to credit cards. They process many millions of transactions yearly for thousands of e-commerce sites and online service providers. Each transaction might expose the different parties to various risks, and containing these risks are crucial.
How Modulai did it
The data was largely structured and tabular, but of various types. The team focused on careful feature engineering and testing with implementability in mind. The preprocessing and modeling flows were built with Python open source libraries, and the proposed implementation is based on various managed cloud services.