Talbert, Jeffrey C

Introduction:

Initially will be specifically working with Dr. Emily Slade and various faculty in running biomedical simulations (no PHI or otherwise HIPAA related data) in various projects from the biostats department. My ultimate role is to provide an interface between the standard analysts and data extraction with advanced computing resources in a variety of biomedical projects.

Project:

Will be variable as I will be working with various faculty as a BMI core employee. My direct supervisor is Darren Henderson and my faculty supervisor will be Dr. Jeffery Talbert (director of the IBI). Initially will be working with Dr. Emily Slade from biostats and collaborating with the HEAL team. Can provide additional details once assigned more projects, but at this point I'm primarily tasked with providing additional documentation/reference for the HPC system, streamline the whole analysis process and project submission for the faculty members who might not be as familiar with the HPC environment, as well as to help keep them up to date on any system/software/library updates that might affect future runs.

Personell:

Dr. Jeffrey C Talbert, PI

Taylor D Smith, Data Management Specialist Sr

Dr. Emily Slade, Biostatistics

Computational Methods:

The primary use of the cluster will be for parallelizing existing code for the sake of large-scale simulations. Much will be in the context of R scripts, but also looking at improving efficiency by developing the same methods in C++ or possibly PyCUDA to facilitate better timing and utilization of resources. The specific methods themselves are in development, I'm still working to help facilitate the basics of the whole process, however, will primarily rely on typical parallelization methods to improve run-time and facilitate further investigations. Will likely also utilize the HPC to streamline/parallelize some machine learning processes, but that is still in development.

Software:

R and python scripts, almost certainly C++ and certainly MPI libraries (using according libraries in virtual environments). Will likely investigate the utility of PyCUDA to make use of GPU clusters as well. Otherwise, that should mostly cover computation environments.

Collaborators:

At this point, this is strictly for UK collaborators as we expand the data science capacities in the IBI and biostats departments. As stated in point B., my faculty sponser will technically be Dr. Jeffery Talbert, director of the IBI, however I will also initially be primarily working with Dr. Emily Slade in Biostats.

Projects