Container technology, and in particular Docker, has revolutionized distributed systems in a very short time. At the Texas Advanced Computing Center, we see the potential for Docker to have an enormous impact on scientific computing as well. Containers enable developers to distribute their applications with all necessary dependencies included in a single file. As a result, the software is immensely more portable and provides far greater reproducibility of results. Additionally, the barrier to entry is greatly reduced, as installation becomes as simple as downloading a file. In this talk we discuss our experiences running Docker on our HPC systems. We present a workflow engine we are building that orchestrates the execution of docker containers from a simple, declarative, structured text file. Finally, we present some current challenges and areas of future development.