“HPC is changing the world. Only through diverse opinions and perspectives can we ensure it changes the world for the better”
Our new series showcases the WHPC Mentorship program mentors. Our hope is that by highlighting these trailblazers—and the sometimes unique paths they followed into the field—other women will feel inspired to envision themselves in similar roles.
Today we talk with Jay Lofstead, Principal Member of Technical Staff at the Sandia National Laboratories in Albuquerque, NM.
What was your path to working with HPC?
I graduated with a BS, MS, and Ph.D. in Computer Science from Georgia Institute of Technology and was a recipient of a 2013 R&D 100 award for my work on the ADIOS IO library. As I was nearing my grad school graduation, I saw three options. First was industry. I had left industry ultimately understanding the different career and employment sectors that researchers have from production software engineers.
Seeing the research side and understanding how it is almost exclusively evidence-based, stable, and long-term focused rather than force of will and profit-motivated helped me eliminate that as an option. For academia, I did not express this as an option to my advisor until too late. I did not properly prepare as a student to get an academic position. When I was graduating in 2010, the market for CS Systems PhDs for academic positions was slim. The national labs were the middle ground offering strong, real-world problems and a strong research budget and focus.
Ultimately, I took the position that was the least well known to me at Sandia and have not regretted it. My position has more freedom than I was likely to get elsewhere and even though I had to start over with a new research project from scratch as a new employee, a disadvantage no one should take without understanding the impact. I was able to establish myself as an independent person within a few years. I would not have accomplished this so quickly had I not been pushed into such an independent position.
Why did you choose a career in HPC?
I didn’t start my career in HPC. I fell into it. My first hands-on exposure to computers was at 9 years old in 1979 and I was programming by the next year. I loved being able to manipulate a tool to do amazing things all through a little effort. I went to undergrad for computer science and intended to go to grad school. I was broke and didn’t know grad school could be subsidized. Instead, I ended up writing programming packages for embedded industrial controls (i.e., the computers that run factories and assembly lines). From there, I went into medical software working on clinician workflow software. When it was time to change jobs again, I had to go to grad school to attain a similar position to the one I wanted to leave.
Once in grad school, I tried learning sciences but discovered I liked actual teaching more than learning and exploring teaching techniques. I also tried software engineering thinking it was about building big systems. It was about understanding the process of building big systems. That pushed me into systems. I was fortunate to get connected with an advisor that had connections with the US DOE national labs and was introduced to working on software infrastructure to help advance large-scale science and engineering. My lifelong love for science was now something I could directly support in my career in computing. From that point onward, HPC has been the core of my computing focus.
What excites you about the future of HPC?
My core research area is data management and workflows. I come from an IO/infrastructure level between applications and storage. I care about how to manage data in ways that make it easier and better for users while adapting to the underlying storage hardware and software. The new advances in persistent devices, such as persistent memory are blurring the lines between what is memory and what is storage. This has the potential to change how applications are written and how users interact with data. With a strong side interest in reproducibility so that we can trust our data, where it came from, and how it was created, we can add stronger scientific rigor to computational science and engineering. With the increased annotation real reproducibility requires, it is another data management problem at its heart.
From a people perspective, I see growing interest in expanding the workforce in any way possible. This has forced many to seek qualified people of all kinds and train them into new careers. In a lot of cases, these are places looking for extra hands-on keyboards that can be harsh and unfriendly. Women in HPC is one organization that I have found that fosters growing the community that will bring new perspectives and opinions that will improve the quality and quantity of our scientific advances. Helping the community grow in positive places that nurture and grow people rather than burn them out and throw them away for the newer, cheaper, more current model is crucial for the long term. Any chance I get to work with anyone to show them the wonderland that is scientific computing on HPC, I take it. I have introduced many to a field that combines both science and computing in a way that lets the individual shine in the best way they can while helping advance the state of knowledge. Helping them navigate the potential minefield of career options to find the best two-way fit is an important part of that. By two-way fit, I mean that the candidate has to be the right person for the job and the company has to be a good fit for the person. Helping people approach an interview as a two-way interview is crucial to make this happen.
What are the top three things you would advise to women considering a career in HPC?
First, don’t quit because things get tough or the situation seems bad or wrong. The situation is probably worse than you realize, but that is not your fault. Instead, seek mentors and friends you can speak openly and privately about things to help you make the right choices. Switching advisors might be best, Changing schools might be required if the program is too small. Changing topics is possible, but always remember that funding is paying you to work on something. If there are issues where you are not respected or it seems hostile, find a way out without quitting. If you quit, they win and you deprive yourself of an amazing career. Ultimately, life may happen or you may realize that it is not the right path for you. If that is true, then go.
Second, understand the difference between a man’s resume and a woman’s (typical case). A woman will tend to not list a skill on a resume until they are truly an expert in it. A man will list a skill on a resume that they are familiar enough with that they believe they can figure out anything they need to on the job. While you may be strongly more qualified, if you do not present yourself in a competitive way, you will not look good. The resume goes through a filter before any interviews. Make it as strong as you can without lying and consider the man approach. I often ask women when interviewing them about the depth of skills and then I know how to interpret the resume. If they know a lot more than the paper says, I give a lot of extra credit. This happens far too often keeping women from attaining positions they are fully qualified for.
Third, this is a lifestyle choice as well as a career. You can be in HPC and work a standard 9-to-5 job or you can grow the job into a monster that takes over your life. With the stability HPC jobs typically offer, you can focus more on your personal life without worrying about potentially changing jobs and moving every couple of years. As with any job, understanding that those that sacrifice their personal life for their work-life are more likely to get promotions and new positions is just part of working. HPC can offer a very satisfying career no matter which path you go down. No matter what kind of lifestyle you choose, you can still work in HPC and help change the world for the better.
What is your most proud accomplishment in the sector to date?
One of my undergrad interns.I managed to bring her in and dumped a completely ill-defined project with no defined path on her. She nearly broke from the experience but bounced back giving me a list of things she didn’t know that she felt kept her from being able to succeed. I helped her get started learning just enough about each topic. She quickly learned what she needed asking questions as she went and then executed with extreme efficiency a small, but complete research project–all within a 12-week internship. As a rising senior, she got a legitimate first author research paper accepted at a top workshop in the field and then presented her work to the audience. Yes, I made an undergrad thinking about grad school present to an audience of 150 of the top people in the field including nearly all of her potential Ph.D. advisors. She was the first talk after the keynote and had no prior exposure to giving a research talk. I coached her through it and she did great. Everyone assumed she was a second or third-year Ph.D. student. She got the DOE Computational Science Graduate Fellowship, one of 26 awarded for the year, and turned down the National Science Foundation Fellowship (2000 offered). She is now about to propose her thesis for her Ph.D. and has grown into a very strong, sharp researcher always asking tough questions without easy answers and handling difficult questions without crumbling. Watching and helping her grow has been immensely satisfying.
I have other technical achievements such as ADIOS, IO500, DataPallets and Stitch-IO, but they don’t seem to have the same kind of impact as bringing someone that was considering an MS degree and a job in the industry as a production software engineer to becoming a strong Ph.D. student looking to work in a research position in HPC knowing that she is going to change the world.