PNAS: Adam DeConinck, High-Performance Computing Systems Engineer

(This post is part of the new round of interviews of non-academic scientists, giving the responses of Adam DeConinck, who works at a company providing supercomputing resources. The goal is to provide some additional information for science students thinking about their fiuture careers, describing options beyond the assumed default Ph.D.--post-doc--academic-job track.)

1) What is your non-academic job?

I work as a systems engineer at R Systems, a company that provides high-performance computing (HPC) resources to the commercial and academic research communities. We own and operate a number of
supercomputers on which we sell computing time, and often heavily customize the systems for whatever application our customer is using at the time. Our customers work in a variety of domains, including meteorology, finance and bioinformatics, but the thing they all have in common is that they have very large computational problems to
solve and need access to hardware, HPC expertise, or both.

I tend to think of my job as being in "research support." I work on a day-to-day basis with other scientists and engineers by helping them to carry out simulations or analyses using unfamiliar tools. I work on the technical side, building and configuring compute clusters for our customers' projects, as well as on customer communication and outreach
to various research communities.

2) What is your science background?

I have a B.S. in Physics from Michigan Technological University and an M.S. in Materials Science and Engineering from the University of Illinois at Urbana-Champaign. I left my PhD program at Illinois after realizing that I no longer wished to work in materials science, and the degree was unnecessary for the types of work I wanted to do.

3) What led you to this job?

Leaving a PhD program is scary: except for a summer internship, I had spent my entire career so far in academia, and the culture in academia makes it difficult to imagine any other life. But while I was miserable in my program and no longer interested in my research, I had built relationships in the community that I did not want to give up--in
particular with my now-fiance, Leigh. Thankfully, Leigh and my family and friends were incredibly supportive and encouraged me to search for work I found enjoyable and interesting. It helped that Champaign, like many research universities, is surrounded by private companies focused on commercializing and supporting research.

My most rewarding experiences had been heavily computational, so when I started my local job search, I looked for work that would play to those strengths. Thankfully Champaign has a long history of scientific computing both at the University and in spin-offs, and I found R Systems.While I didn't have a conventional background (no CS degree), I did join the company with a lot of experience from the user side of HPC. This has proved to be very helpful in working with our diverse customers.

4) What's your work environment like?

We work in a large, open-plan office (no cubes!) with a number of tables
and office chairs scattered about, and a nook with a couch, chairs and a
TV. It's a pretty relaxing environment, especially compared to all the
basement labs in which I've worked. I also do some work in our two
datacenters, both of which are very close to the office.

5) What do you do in a typical day?

On any given day I might be setting up a compute cluster for a new
project; helping a researcher solve a problem they've encountered with
their application; running benchmarks or evaluations using a new
technology we're looking at deploying; or working on automation or
documentation. A given project might last anywhere from a few days to
over a year, so our setup is pretty dynamic.

A lot of the work is very project-oriented, as you'd expect, but we also
do a fair amount of R&D to try to make our clusters run faster and
expand their capabilities. High performance is the name of the game, and
my best days are the ones in which I can make a researcher's life easier
before they've even asked for help.

I also spend some time visiting customers to work with them directly,
and traveling to conferences to speak about HPC and scientific
computing. Those aren't my typical days, but they can be a lot of fun.

6) How does your science background help you in your job?

From time to time it's been useful to understand a researcher's
scientific problem in detail in order to troubleshoot their system.I
don't generally work on their code directly, but a good understanding of
their data and their software can be useful in ensuring their results
are valid as well as performant.

More often, a background in science is enormously helpful in
communicating with the researchers using our systems. It's not uncommon
that working with us is the first experience a researcher might have in
HPC, and speaking a "common language" can make life a lot easier,
especially in getting everything set up right the first time.Our
customers come from a wide variety of domains, and I can't expect to
understand the details of the scientific problem as well as the
researcher; however, I can often understand well enough to anticipate
potential trouble spots or opportunities, and translate the issues
they'll face working on our systems.

7) If a current college student wanted to get a job like yours, how
should they go about it?

The traditional entry into HPC is through a CS degree, but there are a
lot of people who entered from other sciences. I know a number of
physicists, chemists, and biologists who ended up in HPC through one
path or another. Most of them ended up doing some computational work as
part of their "regular" research, and ended up in HPC when they
discovered they enjoyed it. It helps to pay attention to the
computational side of your field, and look for ways to take advantage of
computing to solve research problems.

It also helps to be a generalist, work with people in other fields, and
develop strong communication skills. That's not necessarily encouraged in
grad school where it can be easy to become very specialized, but being
good at working outside your field of expertise can be helpful both
inside and outside of academe in general.

8) What's the most important thing you learned from science?

I think the most important thing I've learned is a deep understanding of
the scientific process. The ability to isolate and test individual
variables, look for possible causes of unexplained behavior, and have a
healthy respect for both the usefulness of data and how easily it can be
misinterpreted, has been useful to me in my HPC projects.

9) What advice would you give to young science students trying to plan
their careers?

Keep in mind that you don't have to work for a University, be a PI, or
even have a PhD in order to do science and work on interesting
problems. The world outside of academia has many places for a person with
research skills and a scientific background. I work with commercial
researchers on a day to day basis, and there are many other jobs which,
like mine, focus on supporting researchers or building their tools.

10) (Totally Optional Question) What's the pay like?

Without being too specific, I'll say that the pay in HPC is comparable
with salaries in the rest of the technology sector, and is therefore not
too shabby.