The Urban Institute chose a cloud-first strategy with AWS to increase the capacity and performance of big data and high-performance computing workloads for 300+ researchers. In this chalk talk, we demonstrate how our team piloted the execution of one of our signature microsimulation models in parallel, designing a new architecture in the cloud that required minimal changes to the original model’s source code. We also show how our team worked directly with AWS to develop open source code that launches powerful Amazon Elastic MapReduce (Amazon EMR) Spark clusters with minimal setup time, and combined it with our Elastic Cloud Computing Environment to allow for a simple Spark user experience.