Transitioning from academia to business, industry, and government used to seem like a black hole into which people and information would disappear. Nowadays the division between the two sectors is not as opaque, as more details—in the form of conference presentations, panels, and written and oral anecdotes—are readily becoming available. Yet because information and personal accounts about industrial careers in the biotechnology and pharmaceutical realm are still somewhat scarce, we asked applied mathematicians Andy Stein of Novartis Pharmaceuticals and Dean Bottino of Takeda Pharmaceuticals to offer readers a glimpse from beyond the event horizon.
SIAM News: Tell us a bit about your respective educational and career paths.
Andy Stein: I earned a bachelor’s and master’s degree in mechanical engineering from the Massachusetts Institute of Technology, then started my career in biotech during a summer internship at Novartis (working with Dean Bottino) while I was pursuing my Ph.D. in applied mathematics at the University of Michigan. After completing a postdoctoral appointment at the Institute for Mathematics and its Applications (IMA) at the University of Minnesota, I joined Novartis full-time in 2009 and have worked there ever since.
Dean Bottino: I earned a bachelor’s degree in mathematics from Hamilton College and a Ph.D. in applied mathematics from Tulane University, then completed a research instructorship at the University of Utah and a National Institutes of Health postdoctoral appointment at the University of California, Berkeley. My first biotech experience was at Physiome Sciences in 2000, after which I co-founded the BioAnalytics Group LLC in 2003; I next joined Novartis for seven years, Roche Pharmaceuticals for two years, and have been at Takeda from 2013 to present. I have spent the last 18 years working nearly exclusively in oncology.
SN: What led you to make the transition from academia to biotech?
AS: There was a lot that I loved about academia, but I couldn’t figure out where I fit. I was hesitant to join an engineering department due to the pressure to bring in grants, and I was hesitant to join a math department because the only pure math class I had ever taken was real analysis, and the work that I was doing—and tend to still do—was very much focused on solving problems and light on mathematical proofs. I had enjoyed my summer internship at Novartis, so I applied for and accepted a full-time position at the end of my postdoc.
DB: During my academic work in mathematical biology, there always seemed to be an introductory paragraph that started with, “Cancer is bad. Indeed, let \(\tau_\gamma\) be a tensor field on the \(N\)-manifold \(\gamma\)….” In truth, I couldn’t really see the connection between my work and explicit improvements to the human condition. I went into industry hoping to find a more direct link between mathematics and its ability to improve the quality and duration of lives for people with diseases.
SN: How can individuals know whether they will succeed in biotech?
AS: The only way to really know is to try it, and I’d highly recommend an internship or industry postdoctoral appointment. I began my full-time career at Novartis as a “visiting scientist,” which basically meant postdoc. It paid slightly less than a standard full-time position, but I could conduct more focused work on scientific problems—much like what I was accustomed to in academia—and was able to publish. Communication and teamwork are much more important in industry than in academia, and deriving novel mathematical results is less so. If you can answer critical questions using high school math and easily explain your answers to the team, everyone will greatly appreciate you. But if you prove a new result or develop a new method that your team can’t understand and that doesn’t help them solve a valuable problem, no one will care.
DB: Academia is great if you aspire to be the best <adjective><adjective><adjective> mathematician on Earth (remove one adjective if you’re a rock star). But if you like applying math to solve important and interesting problems in medicine—irrespective of the math’s “fanciness”—and you enjoy working with large teams on issues that are much too big for one person to tackle, then biotech could very well be for you.
SN: What is your typical day like?
AS: It’s a mix of coding in R, attending meetings (either one-on-one conversations or team meetings), reading papers, putting together presentations, and going for walks to think. Since the start of the pandemic, I have been going into the office about once a week; while there, I conduct mini hackathons with the other pharmacometrician on my team to develop tools that will help the entire department.
As long as students and early-career researchers continue to stay curious, they are likely to find rewarding careers regardless of whether they ultimately opt for industry or academic paths. Figure courtesy of
Tim Urban at Wait But Why.
DB: My answer is similar to Andy’s, but I’m not as good as he is at getting out of meetings. Some days I just go from meeting to meeting, either to mentor junior modelers or lead cross-disciplinary teams toward some shared goal that involves mathematical modeling. Other days I do hands-on tasks like building models, fitting them to data, and reporting the results.
SN: What coursework should graduate students pursue to prepare for a career in biotech or pharmaceuticals?
AS: You should learn to code. Python and R are good choices, but know that you’ll probably have to learn a new language at some point. I knew MATLAB when I joined Novartis and used it for a while, but my group and field moved heavily into R so I eventually made the switch.
DB: I was an anti-statistics snob in graduate school, but that attitude (and the resulting exclusively deterministic skillset) did not serve me well in biotech. All key decisions are made under uncertainty and involve noisy sparse data, interpatient variability, and prediction uncertainty. I had to pick up a lot of stats on my own and wish that I had learned more of it in school—like nonlinear regression, survival analysis, and statistical physics—in addition to deterministic math.
AS: Oh yeah, stats is a good one. Along with all of those analysis techniques, stats also teaches design. Let’s say that you want to create a study to answer a question; how can you do so in a systematic way and account for as many confounding factors as possible? This type of project falls under the framework of “causality” and is one of the hardest topics I’ve encountered, just because of how counterintuitive everything can be.
SN: What do you miss about academia, and what are you glad to have left behind?
AS: I miss being able to submit for publication any interesting results that I find. Now, if I’m working on a drug that has billions of dollars of revenue, uses data from hundreds of patients, and was developed by dozens of people, much more stakeholder management is required in order to publish. I’ve partially addressed this challenge by focusing on the development of new methodologies and trying to use data from drugs that are already in the literature. But regardless, maintaining all of these relationships and following the correct operational procedures can be exhausting.
In academia, I found it quite hard to select a single problem on which to work; I couldn’t figure out how to prioritize or when to give up on something that was too difficult. I like that there is a natural focus in industry — I want to help my team quickly determine whether a drug works and if so, how to optimally dose it and get it approved as quickly and efficiently as possible. Within that framework, I’m free to be as creative as I can when coming up with solutions.
DB: Sometimes I miss teaching and working with students, but those interactions have been replaced by mentoring other modelers and explaining my thinking to non-mathematicians, which I greatly enjoy. The main thing that I don’t miss about academia is balancing my time between teaching and research; I liked both but hated the constant feeling that I was neglecting one or the other at any given moment.
AS: I also miss working with students, but I’ve found that it is possible to get involved in those type of endeavors if they’re important to you. I’ve participated in various collaborations with academia, including the IMA’s Math-to-Industry Boot Camp and a New Jersey Institute of Technology Workshop on Mathematical Problems in Industry. I enjoyed these so much that Novartis has twice hosted a two-week Academia-Industry Hackathon, during which students learned about basic drug development concepts for a few days and then worked on relevant problems.
SN: The ladder to success in academia typically involves publishing papers, receiving grants, and teaching and mentoring students. What is the “currency” for success in the biotech industry?
AS: There are many different currencies for success. Some easily quantifiable measures are salary, job title, and number of direct reports. Other currencies include the drugs that you’ve worked on and advanced, and sometimes the publications that you’ve written. Certain things are harder to measure but perhaps even more important, such as the quality of your relationships with your collaborators. More than anything else, “relationship currency” will probably help you achieve the career that you want. Unlike in academia, there’s no obvious tenure-adjacent goal in industry. This means that it’s up to you to pursue what’s important to your career and life. This is not necessarily easy to figure out, and I’ve shared further thoughts on this topic on my website.
DB: In my opinion, the main currency is the ability to use principled quantitative thinking to impact key decisions in drug programs. There is not a big emphasis on novelty, at least not on being “first but wrong.” You get additional points for working well across disciplines and operating outside of your comfort zone.
SN: Do you have any advice for academicians whose students might be interested in biotech?
AS: In the drug development field, it is not uncommon for people to leave industry and return to academia. Those people are often particularly good researchers and teachers because they understand both theory and practice, and they can be good resources for your students.
DB: Sometimes it’s hard when a bright young person doesn’t want to follow in your footsteps — I’ve been there too! But don’t take it personally; like biotech, academia is not for everyone. Encourage your students to network and seek out industry internships, then give them the time and space to go all in. Shorter commitments, such as weeklong industrial-academic workshops like the Fields Institute’s Systems Modeling in the Pharmaceutical Industry Problem Solving Workshop, can also give students a feel for whether they want to try a lengthier internship.