Early Life and Education
I was born in Hawaii and moved to the mainland when I was quite young. I grew up in the suburbs of central New Jersey, just off of 8A on the turnpike. After completing high school, I attended the School of Engineering at Rutgers University where I majored in ceramic engineering and spent a year and change doing research using molecular simulation at the Interfacial Molecular Science Lab. I graduated in the final class of ceramic engineers at Rutgers; the program was replaced by one in materials science and engineering in the following year.
Ceramic science is a fascinating one, but a field trip to a local glass factory in my final year as an undergraduate made me realize that I really did not want to be an engineer when I grew up. To defer having to go out into the real world to make a living, I opted to go to graduate school instead and enrolled in the materials engineering program at Lehigh University in Bethlehem, Pennsylvania. My hope was to become an electron microscopist, as Lehigh had some of the world’s most sophisticated aberration-corrected scanning transmission electron microscopes at the time (supercomputers of the microscope world, in a sense).
Those plans did not pan out, and after a semester in Pennsylvania, I returned to Rutgers to continue doing research with molecular dynamics simulations.
Getting into Supercomputing
In the four years I was doing my Ph.D. work at Rutgers, I started collecting old UNIX workstations. In the course of playing with them all, I began benchmarking my molecular dynamics codes on them, and this gave me an appreciation for the nuances and performance features of various old RISC architectures. I had also started getting involved in the HPC community online, and I credit my friends (now colleagues) on IRC and Twitter with providing the inspiration and knowledge to consider HPC as a career rather than a mere hobby.
During graduate school I also married my wife (and spent two months doing research out of her father’s tractor shop on the Canadian prairie–an interesting experience!). When I finally completed my dissertation, I made the decision to change careers and pursue my interest in HPC full-time. And, since I had also forced my wife to live with me in suburban New Jersey for four years, I looked for jobs in places that were most un-like suburban New Jersey. Much to my great fortune, the San Diego Supercomputer Center had an open position in user services, and I was able to trick them into thinking I knew enough about HPC to be employable.
SDSC provided me with access to the resources, expertise, and support to develop a strong foundation in high-performance computing, and my role in user services put me in a position to solve challenging issues in a variety of scientific domains. Most notably, the next-generation sequencing industry (largely centered in San Diego) began growing into the realm of HPC at that time, and my understanding of the intricacies of SDSC’s exotic data-intensive supercomputer led me to a number of consulting projects in sequencing.
My analyses of the computational requirements demanded by DNA sequencing caught the attention of a few companies who were technologically ahead of the curve. I was eventually extended a unique opportunity to join an early-stage sequencing startup that was developing a revolutionary new product, and my desire to experience the startup life and be a part of such a rapidly growing industry lured me away from SDSC.
Foray into Genomics
After eighteen months in San Diego, I moved up to Oakland and joined 10X Genomics as a DevOps Engineer (a job which I had no idea how to do). My job involved running a small cluster, maintaining a couple hundred terabytes of storage, and integrating 10X’s software product with HPC infrastructure and cutting-edge DNA sequencers. My heart never left supercomputing though, and shortly after my one-year anniversary at 10X, I returned to the world of supercomputing at the National Energy Research Scientific Computing Center (NERSC) at the Lawrence Berkeley National Laboratory (LBL).
Now at NERSC, I am a part of the Advanced Technologies Group where I try to envision what the future of supercomputing will have to be to meet the needs of scientists. I am one of the resident experts in parallel I/O and new storage technologies, and the emphasis of my work is on parallel I/O system architectures. This includes understanding the low-level hardware details of non-volatile storage media, new software technologies for high-performance I/O transport, optimization points for various parallel storage systems including Lustre, Spectrum Scale, and object stores, and where all of these technologies are headed in the next five to ten years.