Development of a computational and data-enabled science and engineering Ph.D. program

The previous two decades have seen the successful deployment of Computational Science programs in universities across the globe. These programs are aimed at training scientists and engineers to tackle problems requiring interdisciplinary approaches to finding solutions to scientific and engineering problems and the development of new computing, as exemplified by the co-design approach to exascale architectures and applications. Thus, the programs emphasize preparation in applied mathematics, numerical analysis, and scientific computing in addition to science and engineering work relevant to the target application. The rise of so-called 'Big-Data' applications and the use of large data in business decision support and even in computational science workflows like uncertainty analysis are driving a need for training in data sciences. This paper makes the argument that, rather than treating topics in machine learning, statistics, etc. as stand-alone fields of study that students learn as electives, data-science should be an integral part of interdisciplinary training for future researchers. This approach is at the core of the newly developed Computational and Data-Enabled Science and Engineering (CDSE) Ph.D. program at the University of Buffalo. This paper describes the development of the Ph.D. program, the target student audience, and strategies for effectively executing the proposed curriculum.