Friday, October 22, 2021
In modern data science work our stakeholders ask a lot of practitioners, our role expands quickly from model builder to engineer, architect and optimizer when the project requires it. In addition to developing accurate models, projects require the processing of big data, regular retraining, flexible infrastructure, and speedy delivery. Slalom recently completed work to develop a model which predicted customer exposure for mixed channel advertising campaigns. However, the business partners did not want to know a predicted value – they wanted to know the best predicted value and how to get there. Our solution relies on the cohesion of services across the AWS Cloud – including Amazon SageMaker, API Gateway and Amazon ECR – in order to produce a robust and informed recommendation engine which returns optimal budget splits to users in seconds.