Unveiling Risks in Model Migration and Automation

The lack of internal expertise along with the absence of proper documentation for the existing models posed a considerable obstacle in recoding the SAS models into Python and migrating them to AWS. The client knew the importance of comprehensive documentation for future maintenance, troubleshooting, and knowledge transfer but was constrained by prior choices and scant experience in the space. They realized this knowledge gap required external support from a consultant’s specialized knowledge and background to ensure a smooth and efficient transition. The client brought on CG Infinity.

Following a carefully planned process ensuring seamless regression testing and meticulous documentation, CG Infinity’s data scientists partnered with the client to redevelop the models in Python, identifying and resolving any issues that arose. The team then mapped the data while eliminating the need for interaction with the legacy data system through code updates. For the final phase, the models were operationalized using Domino Data Lab with the data sourced from AWS S3. With an eye toward future growth, the model migration provided the client with standardized coding practices, comprehensive documentation, version control mechanisms, and thorough testing procedures. The automation of model scoring significantly reduced manual effort, which had previously been required to run the SAS models. The collaboration with CG Infinity allowed for critical examination that resulted in enhancements and corrections which empowered the client to better serve their business partners.