Faculty Job Interviewing Tips
@jeffbigham
1/21/2019
It’s the time of year when the academic job hunt picks up. By this point, you’ve prepared your materials, submitted them to a bunch of schools and companies, and are eagerly waiting to hear back about next steps. If you’re lucky, maybe you’ve received some phone or on-site interviews. If not, do not despair, in my experience many places won’t have finished up with their first rounds of interview selection just yet. That said, if it hits mid-February, you might consider broadening your search to postdocs or industry positions (which might not have as fixed of an application cycle)… or asking your advisor to be more active in inquiring about your chances.
This document is a pseudo-random collection of tips, which I finally wrote up because I end up telling a lot of people a lot of this over and over again each year. Hope it’s helpful! Send me feedback, since I intend to update it as I’m reminded of or learn new things.
If you know somebody at the place that is interviewing you, talk to them. A faculty member (perhaps, a junior faculty member who was recently your PhD peer), or a PhD student, can know a lot about a place -- how they think, what they’re really after, and maybe what kinds of concerns the particular faculty happen to have when looking for new hires. So, talk to them!
I’ve written previously about “Being the Only CS Person in an HCI Department”: there’s definitely a follow-on article to be written about “Interviewing as an HCI Person in a CS Department”. The first question is -- have they really bought into HCI? I know of at least one highly-regarded CS departments that every year (for like 10 years) has interviewed one of the top HCI people on the market, but has never been able to successfully hire one. Interviewing at that school is very different than interviewing at UW CSE. And interviewing at both places is different than interviewing at CMU HCII.
Your job at the interviewing stage is to find all the great reasons to go to a place. I promise, everywhere has great reasons. Your job is to imagine all of the great collaborations you’ll have with your potential future colleagues. You might not have been to Irvine, CA, or Seattle, WA, or Granville, OH. People live in those places, and there are great things to do there. You might not see your most likely collaborator as the complexity theorist or the opthamologist or the art prof who is really into technology, but for the purposes of your interview your job is to imagine those possibilities. We all understand you’re not going to collaborate with everyone in the university, but we also all know that your collaborators will likely not only be the obvious ones you can tell at a glance. Will you be a good intellectually curious colleague, or will you be a silo onto your discipline?
An overarching goal of an interview is to be curious, to find connections, and to explore. It’s an amazing opportunity to learn about places and people that you haven’t (and maybe won’t) experience again. After all, in the end, you’ll end up at only one place.
Finally, keep in mind that academic institutions, esp., are notoriously bad at interviews. Department heads generally haven’t been there very long, all the faculty may want to meet you to have their say (which leads to long schedules), and, honestly, most faculty haven’t been trained on how to effectively interview someone. It can be chaos. Welcome to academia, in general. Of course, none of this applies to Carnegie Mellon University, where we are highly-skilled professionals.
I’ve generally noticed that phone interviews are becoming more common than they used to be. I think this is generally a way for universities to consider a broader range of candidates, while at the same time “de-risking” -- it doesn’t cost all that much money to bring in a candidate for an on-site visit, but it does take a huge amount of time.
A phone interview lets people get a first cut on (i) how you talk about your research, and (ii) whether you’re serious about their university. For the first one, I recommend practicing at 1-2 minute overview of your research. Maybe brainstorm some common questions you tend to be asked. One amazing part about interviews is that people will take questions in so many predictable but also unpredictable directions. It’s amazing.
For the second part, I think the goal should be to be honest with them. If you really think there is zero chance you’ll go to a particular school, and really aren’t open to even considering it, then tell them that. I had a phone interview long ago, and the place was a lesser known school in a big city, and I had to admit to them and myself that in reality I was pretty unlikely to actually accept a position even if I was lucky enough to get it. Your calculus about where you might consider might change, obviously, based on your other options.
That said, I want to emphasize that so many places are pretty cool. 98% of the time, I would recommend coming up in advance with some of the reasons you would like to go there - the people, the city, family connections to the area, whatever it is. This is especially important for places that might be lower ranked and had a harder time recruiting people in the past -- they might want to know in advance if you’re even interested in theory; it’s not worth anyone’s time if you go to a place without an open mind.
Phone interview can be super weird. Sometimes they are with a single other person, who was specially chosen for having reasonable people skills. Sometimes they are conference calls with a committee, where nobody really knows what’s going on. Either way, your job is to go along with it. Weird conference calls with an unknown number and composition of faculty is a bit unnerving, but it happens. Ultimately, I don’t think it changes your role much -- describe your research in a compelling way.
The “job talk” is super important. In some ways, this will seem unfair -- you’ve done so much during your PhD, why does that one hour matter? Well, it matters for at least a few reasons. The first is that it’s the one common reference point that all of the faculty making hiring decisions will have. The second is that it serves as an imperfect proxy for a lot of things that are hard to otherwise get a handle on -- how might this person be in front of a classroom? Can they explain their research well? Can they get people excited about their research?
On the one hand, this is sort of unfair; on the other hand, you have a lot of time to practice. In fact, probably never again will you have the chance to give a talk so many places within a short time frame. Create a really nice talk, give yourself time to iterate on it, practice it, get feedback. Standard talk advice applies -- I like to boil it down to the audience are your users, how do you design the best talk to convey to them what you and your research are about?
Personally, I like talks that tell a story, provide some structure about what you have done, but also how you think about problems. If it is not obvious, you are not beholden to the existing chronology of your work. Put the pieces (and papers) together in a way that tells the best story. Don’t belabor the uninteresting details. If you’re a Systems HCI researcher, you don’t need to talk about each individual system like it’s a separate thing -- it’s fine to blend them together. Or, maybe you think part of your story is the breadth and so you’ll want to preserve their distinct identifies. I definitely found that putting together my job talk was a great exercise for figuring out how to structure my dissertation.
Ultimately, if you mention something in your job talk, it should be there for a reason. One reason to include things is to describe your research (especially the interesting parts!). But, there are other reasons. One reason is to establish the depth of your work -- you will probably want to go into depth on at least one project. I’m not sure I agree with it entirely, but some folks think there should be at least one slide that goes into enough detail to lose everyone but people in your field. I’m pretty skeptical of the “lose them” part, but I am a supporter of clearly describing technical depth on at least a few slides. This goes slightly counter to the “tell a story” advice.
But, there are other ways to establish your expertise. Long ago, I remember an HCI candidate interviewing at the University of Rochester CS department. She made a point to clearly define some of the methods that she used in her research. For instance, when she introduced a “formative study,” she clearly described what a formative study is, what the goal of one is, and what makes a good one. I remember sitting there at the time that this was pretty strange, going over basic terms, but I was totally wrong -- it worked super well! In the faculty meeting after, people went on and on about how great she had been at rigorously defining her work and what HCI is… I learned a lot from that (one thing I learned is, maybe I should do more of that!).
Not everyone will be from your field, and demonstrating a knowledge of your area (in addition to your specific research) is one way to demonstrate that you’re ready to move from PhD student to faculty. All of us will need, to different degrees, advocate and explain not only their specific work, but also how their area fits into the broader picture.
Personally, I wouldn’t mention specific faculty you’d like to work with in a job talk. This is sometimes awkward, as you never know how people feel about that person, whether that person has been effectively retired for 20 years, or who you inadvertently left out. If you’d like to customize your talk to the faculty at the place you’re interviewing, I would reference them in much more subtle ways, like reference papers by those people or refer to public statements (tweets!) they’ve made. Do this in ways that you could use at any university. If you have other obvious ways to customize to the place you’re at, that can be fun -- e.g., if you’re showing some map application or social media example, try to use the city you’re visiting or people at the place in the example.
For slide design, there are a handful of things that I think are specific to job talks to consider (in addition to making the slides look nice). For instance, I’ve often seen people place little notes about papers that received awards (e.g., Best Papers) next to the references to the papers in the slides.
Another completely unfair, but easy to predict aspect of job talks is dealing with time management. Almost all job talks will be scheduled for about an hour. You can assume that they will start about 5 minutes late, and you can assume that people will expect you to stop speaking about 50 minutes into the talk. These are *almost* universal. What is highly variable, but also predictable in being highly variable, is the department’s culture around asking questions during your talk. What is not variable is that even if it’s entirely the faculty’s fault, they will still expect you to be done by the end of the hour (with some time left for questions). My advice is to plan a 45 minute talk, and practice different endings (perhaps with more or less discussion of future work).
A job talk often will end with a discussion of future work that you’d like to do. One caution about this is that by their nature, the audience will be most likely to ask you questions about whatever it is you covered last. So, this may not be the best time to introduce a cool idea that you’ve not really thought too carefully about. One thing that I think works well is to use the future work as an opportunity to introduce threads of work that didn’t fit into the main story of your talk. For instance, my first job talk was primarily about my accessibility work, but I had a nice thread of work on information extraction and end-user programming -- I had an example for each of these that very briefly described my past work in these areas, and an idea or two about how to extend it going forward. I liked this because it allowed me to introduce some breadth, talk about something I knew about, and while still discussing the future.
I would conclude with a brief summary of your main thesis (of the talk, doesn’t have to be your dissertation’s thesis). I would put whatever you most want questions about on your last slide. Make it easy for the audience to remember the main point of the talk, and make it easy for them to ask questions about the things you most want to talk about!
On-site interviews can be a really great experience. You will likely never again get the opportunity to talk to so many people from so many different backgrounds about the work that you do!
You’ll get asked a lot of predictable questions, but you’ll also be asked questions from curious smart people who don’t know much about your work. Answering them will often be exercises in how to explain your work effectively without the assumptions we all have from knowing our own work and fields well. But, every so often you’ll get a really interesting questions, that might even get you to think about your own work in a different way. That can be super exciting!
Sometimes you’ll get really frustrating questions. When I interviewed in CS departments, I often got asked why HCI belongs in CS. It’s not fair, exactly, that you get asked to defend your field when you are mostly interested in promoting your work within it. Especially, since more traditional/accepted fields within CS do not get asked these sorts of questions. Nevertheless, it’s good to have thoughtful answers to these questions.
In general, the goal of an interview is to present yourself as being open and excited about the various research and other opportunities available to you if you chose to come to a particular university. This is good because it’s also the perspective I would advise you to take. Be curious, be collaborative, and be friendly. These could be your future colleagues, this could be your future university, and this could be your future town!
Interviews are a great chance to start to form an opinion about the university, and learn more about how it operates. This might not be immediately useful, but as you advance through your career it’s really important (and useful) in my opinion to get a sense of how different departments, schools, and universities operate. For instance, who holds what power in the department you’re interviewing with? In some places (like CMU), the department head holds a lot of power and largely controls her own budget; whereas, as other places (like Rochester), the department chair has relatively modest budget and the dean of the school takes a much larger role. The variations on how things work are endless -- these details mostly don’t matter, but they are interesting, and maybe impact things like who you will eventually end up negotiating with.
Interviews can be incredibly grueling. Take care of yourself! Relatedly, my advice is to not worry about the occasional extra expense. For the most part, nobody will care if you incur a reasonable extra expense, even if it’s to just make your life a little easier. If Delta is going to charge you $50 extra to get on the earlier flight home, after your long interview, pay the $50 and charge it to the university. I usually wouldn’t change flights and incur that fee, but I would totally do that if it’s in the middle of a grueling 18 interviews in 2 months, haha. Your experience could vary, but in general I’ve found computer science departments during interviews to be pretty lenient.
Interviewing is expensive, in that the assumption had been for a long time that you would just pay all the travel fees and then get reimbursed some time later. I think people are becoming more aware that this is a huge burden and can have a discriminatory filtering effect. Often times you can pay for at least travel expenses through the university’s travel agent and not have to pay upfront, so explore that option.
If you haven’t yet, this is a great time to sign up for airline rewards programs :)
Listen to someone other than me about what to wear, but if you are male-identifying in a CS department, I don’t think you can really go wrong. I probably wouldn’t advise people to wear a full suit, as it seems a bit too formal, but, other than that, maybe just dress up one level from where you usually do. I don’t even care if people wear a t-shirt, but if you’re going to wear a t-shirt make it a nice/cool one. I generally wore dress pants, a dress shirt, and a jacket - kind of boring but definitely acceptable. Sometimes I’d wear a tie. I’d like to believe this advice that mostly anything goes applies to candidates who are women, but advice for women on what to wear is, I think, generally agreed to be more complicated -- Jean Yang’s blog post is a reasonable place to start.
Interviewing is a weird, exciting, grueling opportunity, that you’ll likely only do a very small handful of times in your career. Take care of yourself, be open to the opportunities before you, and try to enjoy it.
This page and contents are copyright Jeffrey P. Bigham except where noted.
Blog posts are not intended to be final products, but rather a reflection of current thinking and/or catalysts for discussion, like tweets but longer.