Too often, data scientists build potentially high-impact models that never see the light of day due to deployment obstacles. At Convoy, data science is at the core of its trucking network product, so building a robust, frictionless platform to deploy models is critical to the company's success. To minimize the code needed to transition from training a model locally to deploying the same model in production, the company adopted Amazon SageMaker to push various models to production-ready endpoints. In this talk, Convoy's data science director, David Tsai, reviews Convoy's Amazon SageMaker architecture and highlights how it has enabled them to drive meaningful impact.