Reflecting on Two Years of Professorship

Reflecting on Two Years of Professorship
Colin Raffel
July 1st, 2022
colinraffel.com/blog

As of today, I've been an assistant professor of computer science at UNC for two years. I've been reflecting on aspects of being a professor that ultimately led me to struggle at times. Specifically, I think there is a confluence of factors (the extent and effect of which I hadn't anticipated) that can make it really easy to experience impostor syndrome and self-doubt. The aim of this blog post is to share some of those reflections so that people interested in being a professor might know a little better what to expect and so that other early-career faculty might feel validated if they have experienced similar things. As a caveat, my perspective may be somewhat specific to my (sub-)field as well as my path to professorship (I spent four years as a research scientist at Google Brain before starting at UNC). In addition, I'd like to note that while this blog post is mainly focused on things that led me to struggle, I have loved being a professor — on the whole, it has been incredibly rewarding and fulfilling.

   

A recipe for self-doubt

As a professor in computer science, one of your main roles is to build and run a lab. In this role, you are effectively the leader of a small team. I had anticipated that this would involve a great deal of responsibility, management, and leadership. What I hadn't anticipated was that you no longer have a path to direct feedback. Unlike being a PhD student or industrial researcher, you no longer have an advisor or boss. To some degree, your department chair is your boss, but the relationship between a department chair and a junior faculty is typically dramatically more hands-off. In reality, the people who are best suited to give you direct feedback are your lab members – after all, they know best how you are functioning as a leader. However, there is a fundamental power dynamic that is difficult to overcome since at the end of the day your lab members are your reports. As many times as I encourage my lab members to give me honest feedback, I know that it may be difficult for them to be completely forthcoming. Ultimately, the most important feedback you get is very indirect or long-term - for example, whether a grant proposal gets accepted, whether a paper gets published, or whether your students are graduating and finding good jobs.

Something I had correctly anticipated about being a professor was that I would be busy. Between teaching, applying for grants, meeting with students, and many professional duties, it is easy to fill all of your working hours (and beyond). In reference to this, the primary piece of advice that I've been given repeatedly is that I should learn to say no to things. While this might be helpful, no matter how many times you say no you still have a very limited amount of time to complete a huge number of tasks (not to mention that people don't like it when you tell them no and saying no can also lead to missed opportunities...). For me, the impact of this has been that I have to complete tasks as fast as possible. In practice, this sometimes means doing the bare minimum amount of work on something and hoping (or trusting) that it is good enough. This has been difficult for me because I value being thorough and spending the time to do the best job I possibly can.

Additionally, for many of the parts of your job that are keeping you busy, you have virtually no experience doing them. For example, I personally had never written (or even read) a proper grant proposal until I started my professorship. It is not unusual for a junior faculty to have never taught a class or formally mentored anyone. While learning new skills is a part of any job, as a new professor I think there are an especially large number of new skills to learn with an especially limited time budget.

Overshadowing all of these factors is the reality that as a professor you are responsible not only for yourself but for the success of all of your students. In other words, if you make a mistake or fail, it doesn't just impact you anymore — there is a group of people who are directly relying on you to do a good job. For me, this leads to a great deal of pressure to succeed.

Combining all of the above leads to a particularly potent recipe for self-doubt. To build a successful lab and support your students, you must perform many tasks (more than you have time to do thoroughly), most of which you have limited or no experience doing, all while receiving limited or no direct feedback. I felt this self-doubt most strongly last summer when I was writing my CAREER proposal. I was writing a solo 15-page grant proposal for the first time, attempting to credibly predict what research my group would be doing in five years and making the case that it would be a sure-fire success. I knew if I didn't get the CAREER or other funding, I wouldn't be able to continue to support all of my students. If I spent too much time on the proposal, I'd ultimately be shirking other responsibilities. I had pretty severe writer's block (which, thankfully, is rare for me) and struggled a great deal to get the proposal done at all. I also recognize that, as a part of demographic groups that are overrepresented in computer science (a white cis male whose parents were both professors), I have a great deal of privilege and implicit/explicit signals that I can succeed. I can only imagine that these factors are much more likely to lead to imposter syndrome in people who come from underrepresented groups.

   

The good news

How can you survive as an early-career professor given all of the above? I have found a few things to be helpful. The first is that not-so-early-career professors seem to be particularly generous when you ask them for advice. I have had great luck cold-emailing professors to ask them random pieces of advice. For example, a few months ago I noticed that a successful professor had a relatively small lab so I emailed him to ask what his strategy was for choosing a lab size and he emailed back a multiple-paragraph response. When I get the chance to chat with someone more senior than I am, I often take the opportunity to bounce ideas and questions off of them as a way of getting feedback on my own strategies. Some of the useful tips I've received include avoiding staking your self-worth on external outcomes, learning to compartmentalize so that work stress doesn't become life stress, and remembering the reasons you became a professor in the first place (because you are passionate and excited about your research area, because you love mentoring, and because you want intellectual freedom — or something like that).

In addition, something I have frequently been told by more senior professors is that they are still flying by the seat of their pants and no one really knows what they are doing, which is both encouraging and frightening. Beyond other professors, I have managed to get my students to give me feedback by badgering them until they tell me something negative. I also started adding a link at the end of my talks to a form that allows attendees to give me anonymous feedback. Mechanisms like these provide small but meaningful ways to get more direct feedback.

I have also found that after I struggle through something, it gets easier the next time I have to do it. For example, after the painful process of writing my CAREER proposal I quickly found that I had an easier time writing other grants. I think the fact that I was able to eventually complete and submit a proposal helped prove to myself that I was capable, irrespective of the outcome. In addition, on my best days I try to remind myself that the reason I managed to become a professor at all was that I was good enough at certain things (writing and publishing papers, giving talks, etc.). This helps give me some reassurance that I will do an ok job even when I need to complete a task with a super limited time budget.

Overall, I am happy I am a professor. It has been a struggle at times, but it is hard to imagine a job that is as intellectually stimulating, exciting, and fulfilling. If you are a fellow machine learning-adjacent junior-career professor who has had similar experiences and you want to commiserate or brainstorm ways of supporting each other, please reach out to me.

Thanks to Natasha Jaques and Emma Strubell, who provided feedback on this blog post.

formatted by Markdeep 1.03