Jared Workman Email & Phone Number

Grand Junction, Co, Us

San Francisco, California, Us

Lead a team of high performance computing engineers that installed, automated, and optimized all software for all of Rescale’s customers. Our speciality included numerical simulation software and data analysis tools as well as ML and AI packages ran on a diversity of cpu and gpu hardware. Primarily worked with Azure and AWS Cloud.Highlights: I led a diverse and high achieving team of high performance computing engineers that were responsible for all software on the Rescale Platform including but not limited to the installation, automation, deployment, and optimization of simulation software for Rescale customers across all supported verticals including, Energy, Defense, Manufacturing, and aerospace design. In all, our multi cloud application suite increased from 1000 to over 5000 under my tenureManaged multi cloud High speed Network optimizations for various fabric types at Azure, AWS, Oracle, and Google to ensure peak performance of cloud resourcesWorked extensively with Azure Cloud and AWS using the HBV2, HBV3, HBV4, and EFA skus to enable hundreds of millions of corehours of simulation time to be run efficiently in the Cloud.Developed software installation and deployment pipelines using Linux Bash and Windows Powershell based scripting in Windows and Linux environments to automate deploys of software to Azure and AWS Cloud. Ensured the stability of 1100 products and 5000 total versions over 4-6 cloud providers, 20 distinct CPU architectures, 5 distinct GPU architectures, 6+ network interconnects, and various flavors of MPI on RedHat, Debian, and Windows systems.Through strategic planning and development of processes and procedures, installations had a direct impact on sales growth over 1.5x-2x year over year for 4 years and an overall reduction in delivery time of 50% in the last year.Managed the operations and interface between HPC and international sales teams. Acted as technical project manager for HPC

San Francisco, California, Us

Responsible for the installation, optimization, and maintenance of HPC software on AWS, Azure, Google, and Oracle Cloud. Applications ranged from fluid solvers to molecular dynamics. Optimization included tuning mpi flavors (OpenMPI, MPICH, Intel MPI, Platform MPI) to various fabric interconnects. Software installs ranged from prebundled packages such as the full ANSYS suite to compiled softwares such as Amber, OpenFoam, etc. Further responsibilities include managing the maintenance of and integration tests for Rescale’s top 80% revenue generating software. Also responsible for managing the HPC team’s workflow using Asana and SFDC.General skills involved extensive scripting using bash and python to configure and optimize software for HPC cloud deploys. Experienced with Slurm, Docker, Singularity, and PBS.

Grand Junction, Co, Us

I was a tenured professor of physics at Colorado Mesa University. While at my previous position I:Designed, built, and maintained a 24 core AMD linux server and a 40 core Intel Xeon Linux server which I have used for research and training students to learn scientific programming and data analysis.Secured a $250,000 scientific observatory for a cost of $30,000 by negotiating with the United States Air Force, my university, and a private philanthropist. https://www.usafa.edu/research/research-centers/astronomical-research-group-observatory/Designed and had built, a second $100,000 observatory and secured funding through a private donor.Ported scientific codes to Python for the open source community (http://cococubed.asu.edu/research_pages/sedov.shtml).Trained students to develop their own fluid dynamical solvers and how to manage, analyze, and visualize the data output as well as how to validate the codes.Ported a relativistic hydrodynamic/radiative transfer code used to synthesize gamma ray burst afterglow spectra from IDL to Python. This code originated from work I personally developed during my doctorate and is one of the most respected tools of its area in the world. Developed and taught a course in computational physics based on the Python programming language and using packages such as NumPy and SciPy.Maintained active collaborations with groups at several universities around the country and one in Switzerland which have resulted in 11 papers (all in top 5 impact factor journals) during this position. Attended and presented scientific research at several conferences.Taught 16 separate courses ranging from algebra based physics to quantum mechanics and computational physics.Mentored over a dozen individual research students with strong initial post graduation employment results.Taught over 1000 students with high evaluations.Received the rank of "Exemplary" faculty on several occasions.Worked in public outreach.

My work at the University of Rochester was a two fold appointment. I worked on the simulation of reconnection in solar flares using both a finite volume/Reimann based magnetohydrodynamic (MHD) code and a particle in cell (PIC) plasma code.The goal of my research was to tie together the macro-scale fluid dynamics of solar flares using the MHD code to the micro-scale particle/shock behavior using the PIC code.The MHD code was based on C++ and the PIC code was an object oriented Fortran code. I successfully implemented new physics, data structures, and optimized the MPI output of the PIC code. Both codes were run at supercomputing centers using several million hours of run time on the systems at NERSC https://www.nersc.gov/. I was co investigator on an annual allocation of approximately 5 million cpu hours during my time at Rochester and maintained this status for 5 years afterwards.To perform the research I learned batch scripting to automate the output and transfer of large data sets between scratch directories and permanent storage. I also learned to use shell scripting to parse the analysis and visualization of the data into manageable chunks. My work resulted in the publication of several papers, the mentoring of a graduate student who was awarded his doctorate, the inclusion of robust new physics modules in the existing codes, and novel data analysis codes such as individual particle tracking and decomposition of large data sets into useable subsets. These data analysis codes and modifications to the existing codes remain in use by researchers in the field today.

While at JILA I focused on two, separate projects. The first project was the study of how magnetic fields effect the transfer of angular momentum in astrophysical accretion disks. In this area of my work I modified and ran parallel simulations of finite difference and finite volume, static and adaptive grid based magnetohydrodynamic (MHD) codes on massively parallel supercomputers. I generated algorithms to subdivide the data output into manageable chunks and produced analysis codes and visualization codes to study the output. I also added new physics modules to the existing codes such as vertical stratification, resistivity, and external forcing effects. I coded primarily in IDL, C, and Fortran.My second project was one I created from scratch. I generated a semi-analytic formulation by which to model the relativistic evolution of a hydrodynamically expanding fireball and the subsequent radiative transfer of photons propagating within it. I then, along with a collaborator with whom I am still working, implemented a numerical code in IDL to solve the semi-analytic formulation. The result of my second project was a numerical framework by which to model the time evolving spectrum of light from a gamma ray burst afterglow. Previous modeling required lengthy reductions of fully four dimensional relativistic hydrodynamic codes taking hundreds of thousands to millions of hours of cpu time. The result of this research effort was a simplified scheme which produced scientifically equivalent results using only hours computer time. This work has continued to evolve for the last decade and has resulted in a scheme which is respected and used world wide in the gamma ray burst community. I am currently experimenting with Markov Chain Monte Carlo algorithms and machine learning adaptations to further refine and speed up the code. I have since had the code ported to Python.

As a teaching assistant my primary responsibility was the bookkeeping of the grades of some 300-600 students a semester. The challenge lay in properly linking several disparate sources of input from many different platforms to individual students properly. I used a combination of Microsoft Excel and Linux shell scripting to accomplish the task.

My research involved modeling the the thermal response of satellite hardware using IDL, developing image filtering/manipulation routines and generating data analysis routines to examine sounding rocket solar images.

I taught remedial mathematics, college algebra, and trigonometry.

Phd, Astrophysics

Ms, Astrophysics

Bs, Physics & Mathematics (Minor)

Ba, Psychology