John LaGrone, PhD

HPC Research Solutions Architect
Southern Methodist University

SMU Email: jlagrone(-at-)smu.edu
Personal Email: john(-at-)lagrone.me
Github: jrlagrone

Summary

Broadly speaking, my interest are in applying mathematical techniques and high performance scientific computing to a variety of interesting problems.

I am currently a HPC research solutions architect at Southern Methodist University (SMU), where I assist on a variety of research projects, provide general HPC support, and help administer two HPC clusters – a primarily CPU based cluster and a NVIDIA DGXA100 SuperPOD. I have fully customized and deployed ColdFront to transition to project based accounting and provide a (mostly) self-service platform for HPC accounts. I actively participate in on-going ColdFront development efforts and co-lead the Core Architecture working group. I am responsible for installing and maintaining most user facing software on SMU’s HPC systems – ranging from packages managed using Spack to hand-tuned compilations and customized implementations. I manage usage metrics using Open XDMoD, assist with Slurm management, and manage customized appplications for Open OnDemand.

Previously, I was a senior computer vision engineer at Scandy, a company in New Orleans specializing in 3D scanning. I was primarily responsible for maintaining and improving existing algorithms as well as developing new mathematical methods for improving 3D scanning. You can view a scan of our late Boston terrier, Lafitte, on Sketchfab

Before that, I was a postdoctoral fellow at Tulane University where I worked on the modeling and simulation of passive and active fibers in viscous flows. The applications range from understanding the dynamics of passive fibers such as actin filaments in flows to oil spill remediation using flagellated bacteria.

As a graduate student at Southern Methodist University, I worked on developing absorbing boundary conditions for wave propagation problems. As part of this work, an open source software library was released to implement these boundary conditions for Maxwell’s Equations. I am continuing to work on this, with the goal of extending the work to elastic waves, which may have applications in seismic imaging.