Three reasons why I chose industry research over academia

-Alex Lu

After a long five years, my stint as a PhD student was finally reaching its end – and that meant I needed to hit the job market. As someone who was graduating from the Department of Computer Science at the University of Toronto with a specialization in computational biology, I had some uncertainty about what kinds of positions I should be applying for. I was at the intersection of two fields with drastically different career trajectories. In the life sciences you are typically expected to spend a few years as a postdoc, but in computer science it’s not uncommon to apply for faculty positions right out of the PhD. I did know that I wanted to stay in academia, so I decided to apply to assistant professor postings first to see if I could successfully convince search committees that I was really more of a computer scientist, and then I’d fall back on postdocs if my search was unsuccessful. What I didn’t know was how much my preferences would evolve through the job search: despite being offered a Canada Research Chair position that would have come with up to a million dollars in federal funding, I ended up choosing an industry research position at Microsoft Research (MSR). What influenced my decision to choose industry over academia?

When I first entered the job market, I had three main ideas about academia that made me believe it was the only option for me. First, I thought it was more accessible. Universities are usually progressive, and each has their own accessibility or disability services department. Even though accessibility legislation exists, I always thought that the expense of hiring interpreters would clash against corporate goals of profit, so I assumed that companies would try to scrape by with the bare minimum of accommodations. Second, I thought it was the only place where I could pursue an independent research agenda. I do a lot of basic research, and I strongly believe that scientific research should be used to enrich everyone’s lives and be accessible to everyone, not held as secrets in a private company. Third, I thought it was the best opportunity for mentorship. In academia, you are required to mentor students. As someone from an underrepresented background on multiple axes, I wanted to make sure that other underrepresented people had the same opportunity to benefit from the academic system as I did. 

So naturally, the majority of my applications went out to universities. I sent out a total of 32 applications to research-intensive institutions globally. I made one exception — I sent out an application to MSR as the sole industry position I applied for. The only reason I sent out the MSR application was because a colleague had transitioned to a position there, and he sent me a Twitter DM inviting me to apply. I was initially resistant, but he reassured me that the application would be no additional work; to apply I just needed to submit the same research statement and CV that I was sending off to academia. I figured I had nothing to lose by applying, and sending off my application took less than thirty minutes on their online portal. 

A few months after I sent out my applications, I starting hearing back from departments. Being a fresh PhD graduate with no postdoctoral training, and with the pandemic causing hiring freezes, I was pessimistic about my prospects — so I was surprised to learn that I had scored several interviews at institutions in the United States, Canada, and Europe. What surprised me was that at the majority of institutions I was interviewed by life sciences departments, while the computer science departments mostly turned me down. I was expecting the opposite given the standards for postdocs in both fields, but it turned out that many life sciences departments were excited about interdisciplinary research and aware of field-specific nuances in training. The majority of institutions where I interviewed were incredibly warm and inclusive, and excellent on access: one consulted with their access department ahead of time and offered me 1.5x the time on the screening interview to account for interpretation delays; another institution’s search committee greeted me in sign language. I think my positive experience may have been influenced by me explicitly identifying as Deaf and queer in my research statements; some institutions may have self-selected out at that point, leaving me with only the progressive departments. 

My interview at MSR came later than most of my interviews in academia, so they had to beat a pretty strong impression as well as my natural resistance to the idea of industry research. So how did they do it? What I found is that MSR systematically challenged each of the misconceptions that I had about academic versus industry research. In doing so, they exposed faults in the academic system. While I was aware of these faults, I always considered them nuisances that I had to accept to join an otherwise principled system. MSR offered some better alternatives, and made me realize that these were things I did not actually have to put up with.

First, on the accessibility front, I was taken back to learn that Microsoft had their own in-house ASL coordination team. During my academic job interviews, I was mostly interviewing at departments that while open-minded, had never employed a Deaf faculty member previously. I expected this would be the case; Deaf STEM PhDs are still rare due to the sheer amount of systematic barriers, so I naturally accepted that I would need to do a lot of explaining and legwork on my needs. I had already devised a strategy to minimize the impact of this inexperience on me: I told each of the departments that they should enlist my current academic interpreters at my PhD institution, so I would not be penalized as departments scrambled to find potentially less-experienced interpreters without being aware of the pitfalls. In contrast, the ASL coordination team at MSR directly reached out directly to me. They were totally on board with my plan, but they also made me aware of their own services. They told me that they provided ASL interpretation for about 40 Deaf employees globally, so they already had a system in place for coordinating interpreters. In fact, their in-house ASL coordinator was actually Deaf herself, and was a certified DI — I had an opportunity to chat with her, and we discussed her plans for building a community for Deaf and hard of hearing people at the company. 

This was the first thing I realized MSR could do better than most academic departments: they were able to bring institutional knowledge on accessibility. I’ve always considered myself fortunate, because as a Deaf person, I’ve rarely had to “fight” for my needs — I’ve mostly worked with accessibility departments and academics who were open-minded and interested in accommodating me. But even when working with institutions that are willing, I still have to allocate a proportion of my energy to explaining who I am and what I need to those who have never worked with Deaf people before. I had always considered that energy tax to be inevitable. Microsoft challenged that belief. While I still have to do some explaining — for example, I’m the first Deaf person they’ve hired in a research role (as opposed to sales or engineering etc.), so there are still specific nuances that come with that — for the first time, I could consider what else I could do with the energy that I would normally expend on defending my existence. 

Second, on the research front, I learned that MSR gave full research autonomy to their researchers. Prior to interviewing at MSR, I mostly knew about industry through informational interviews with start-ups and smaller biotech companies. I was not impressed; in addition to keeping my research scope constrained on what would be immediately beneficial to the company, I would not have the opportunity to publish and disseminate my research to the public. What my interview at MSR taught me was that industry research actually occupies a wider spectrum than that. The researchers there explained to me the structure of the company: while Microsoft itself has its own research team that does work on more product-orientated research, MSR is considered an independent entity. While it receives funding from the parent company, the researchers pursue their own research agenda, publish almost all of their research publicly, and maintain active ties with academic institutions. Finally, I was also very excited about how the research group was set up: unlike a traditional academic department, which is stratified by discipline, MSR is highly interdisciplinary, and you see social scientists, mathematicians, biologists, etc. on the same floor every day, instead of being hosted in a different buildings across campus. As a highly interdisciplinary researcher, I was excited about just how much my research would branch out in an environment where I was in constant contact with people of different research backgrounds than me.

Essentially, my new position looks similar to an assistant professor position, with some key differences: the biggest is that there is no grant-writing or mandated teaching involved. There is also no tenure involved. For me, all of these things were appealing. I had always viewed the grant-writing side of academia as a necessary evil to keep the research churning, and the prospect of just not having to write grants was mind-bogglingly exciting. While I enjoy teaching, I enjoy it with highly-motivated students who are there to learn, and I don’t like the administration aspects of more routine courses where many students are just there to check off a requirement on a degree. As for tenure — while the end-prospect was exciting, I was concerned about how the demands of the tenure track might change my values and research philosophy. One of the things I expressed while interviewing with academic departments was that I didn’t want to sign up for a school where publication requirements for tenure were too demanding. I felt that stress of churning out papers might trickle down to my students, and I wanted to be someone who would hold space for my students to learn and explore their own interests instead of expecting them to be productive to bolster my own portfolio. 

Third, I realized that there is still so much mentoring that can be done outside of formal academic structures. One of the disadvantages of the MSR position for me was that I would not be able to build a lab and mentor students through a graduate degree. I initially considered this a serious demerit that would clash with my goals of fostering underrepresented students through the academic system. However, the other benefits of the MSR job made me think about alternative ways I could achieve this goal. I realized that the autonomy baked into the job would still give me a lot of opportunity to do good. For example, each researcher at MSR has the opportunity to hire an intern from a graduate program every year, and I consider this a way to give students opportunities and expose them to new research interests; I plan to keep the scope of my hiring wide and will be looking at outreach from institutions that serve underrepresented students, like HBCUs. Similarly, a lot of MSR researchers take on voluntary academic supervision or service appointments: some serve as diversity chairs for conferences, and many sit on PhD committees. I’m already discussing co-supervising some postdocs, and opportunities for more machine learning education at Gallaudet. I would say that not limiting my mentorship options to the boxes that academia provides for me may foster more creativity, and I’m looking forward to how I carry this out in the future. 

Overall, my interview at MSR left a major impact on me. And this meant I had a very difficult decision to make: my offer at MSR came at the same time as an offer from a major Canadian university. I remember sitting at my desk with both offers side-by-side, and thinking how my offer from academia was everything I had ever wanted. Even when I was a child, my parents had always encouraged me to go into higher education, because they said that with my disability, public institutions would be willing to accommodate me but not private companies. In some respects, I had been groomed into viewing academia as the place for me, and I also feel like this story is true for many other disabled people. I think this is a motivating factor behind much of the activism around accessibility in higher education, because there is a dissonance between the way we are trained to see academia as a sanctuary, and the way it actually is in practice. But in the end, my industry offer won out. It promised a brighter future, without many of the things that I had settled for as a matter of “this is just the way things are” in academia. While we will see if those promises bear fruit, for me, the risk is worth it.

By day, Alex Lu is a computational biologist whose research focuses on artificial intelligence that can “teach themselves” biology in large-scale microscopy datasets through puzzle-solving and interaction. He holds a PhD in Computer Science from the University of Toronto, and will be starting as a Senior Researcher at Microsoft Research in September. By night, he is a Deaf-queer community organizer. He previously served as a board director for the BC Rainbow Alliance of the Deaf, the frank theatre company, and OPIRG-Toronto. His work as a journalist focuses on the intersection between disability, queer, and racial communities.