By Christine Klymko
As you near completion of your dissertation in applied/computational mathematics, one of the most common questions (at least, a question that I was often asked) is, “industry or academia?” The tradeoffs between the two are assumed to be clear: a high salary and some geographical choice versus research freedom and conference travel. Based on advice from my advisor and two positive summer internships, I chose a third path and became a postdoc at a national laboratory.
I am currently finishing my postdoc in the Center for Applied Scientific Computing (CASC) at Lawrence Livermore National Laboratory (LLNL), where I have spent the past two and a half years. My research focuses on the use of numerical linear algebra and graph theory to develop and apply methodologies for analyzing real-world complex networks. In many ways, my postdoc at LLNL is a hybrid between an academic and an industrial position. The work done at CASC ranges from the application of mathematics to solve difficult problems at the lab to the development and publication of fundamental research. While research staff in CASC need to ensure that they have funding, in general the range of funding sources is large enough to allow people the ability to choose projects that fit their research interests and publication goals. Postdocs, however, are encouraged to spend the majority of their time on publishable research. My time is currently split between three projects: (1) a study on how to measure the importance of computers on the lab network to help make operational decisions, (2) a fundamental research study on random walks in temporal graphs, and (3) work with external collaborators that uses machine learning to classify malware. The last project is done in my “25% time,” time given to every LLNL postdoc that allows them enough freedom to pursue their own research interests, begin to establish credentials as an independent researcher, and successfully launch their career. This “25% time” attracted me to LLNL as I was applying to postdoctoral positions.
Any lab-funded research must be geared towards advancing the lab’s mission, which, in the case of LLNL, is the development and application of science with the goal of strengthening national security in the United States. The majority of projects at the lab fall within one of the main mission areas (which include biosecurity, energy, intelligence, and science), although some projects are also funded by various sponsors and clients of the lab (mostly other organizations within the U.S. government). As a result, postdocs and staff researchers at LLNL do not typically have quite the same breadth of freedom in terms of research direction as those in academia. However, I have not personally found this restrictive. As an applied mathematician, I enjoy the motivation that comes from developing math to solve specific problems in other domains. There are plenty of opportunities for fundamental mathematics research within the scope of the laboratory mission, and my work at LLNL has given me many occasions to pursue interesting research.
The main differences between being a postdoc at a national laboratory versus one at a university (to the best of my knowledge, given that this will be my only postdoc) are twofold. First is the relative lack of students and classes in the laboratory environment; during the academic year, few students are at the lab. Although many undergraduate and graduate student interns are present during the summer, postdocs have no requirement to teach or mentor. However, CASC postdocs are encouraged to interact with the interns working on their projects and even take on interns of their own. Last summer, I mentored one graduate intern. The second main difference between a university postdoc and a national laboratory postdoc concerns the dissemination of research. While working at the lab, any paper that postdocs publish or slides they make for a talk must go through a formal (though relatively painless) “review and release” before public presentation. This is because sensitive and classified work is conducted at the lab, and it is necessary to ensure that any released materials don’t contain such information. Additionally, some research at the laboratory cannot be published due to the nature of the work. However, it is not required that postdocs obtain a clearance, and in CASC we are encouraged to spend the majority of our time on unclassified, publishable work. The biggest consequence of this “review and release” process has been that I finish presentation slides before leaving for a conference rather than in the hotel room the night before, which is actually less stressful!
As I approach the end of my postdoc at LLNL, I am preparing to transition to a staff researcher position here. I have also known postdocs who have transitioned to successful careers in academia or industry, as well as others who have stayed within the national lab system. Thus, a national lab postdoc is a great way to advance research and skills without closing doors on any specific career path. I personally feel that my postdoc has exposed many options for me and helped me to prepare for and decide upon the next steps in my mathematics career.
This article is reposted with permission and light edits from LA Digest, the monthly newsletter of the SIAM Activity Group on Linear Algebra.
Christine Klymko is currently a postdoctoral researcher in the Center for Applied Scientific Computing (CASC) at Lawrence Livermore National Laboratory.