Because many Grid-enabled workflow applications (including the next release of EMAN) are driven by high-level scripts, we have begun research on tools to support Python and other scripting languages. In the context of the Grid, this may include automatic generation of workflow from a Python script describing the application. At a finer granularity, we also plan to work on automated distribution and alignment of data on parallel machines and clusters. However, solving all of these problems would, we believe, involve type inference techniques. We have made significant progress in type inference techniques for scripting languages such as Matlab in collaboration with the telescoping languages project, and we hope that the same technology, with some modifications, may be applied to Python for Grid problems.