Sunday, December 9, 2012

How to get a faculty job, Part 1: The application

This is going to be the first in a series of three blog posts on getting a faculty job in Computer Science. Part one is about applying for the job. Part two will be about doing interviews. And part three will be about negotiating the offer and making a decision.

I did my faculty job search back in 2002 after finishing my PhD at UC Berkeley. Back then, academic Computer Science departments were hiring like crazy and the number of job openings far outstripped the number of highly-qualified applicants. I ended up with something like a dozen interviews, and also interviewed at IBM Research (both coasts), HP Labs, and a little search engine startup called Google. (I regret not having interviewed at Microsoft Research, but at the time I was dead-set on an academic position and had a hard time seeing myself working at MSR.) I got offers at all of the industry places and several of the universities; and ended up taking a faculty job at Harvard.

The process of getting an academic job is tremendously painful and takes months of effort. Faculty job applications are usually due in December or January, interviews happen around March and April, and job offers made in April and May. Before summer break most job applicants will have their position sorted out and know where they will be heading in the fall.

The job application itself usually consists of five components: Your CV, a cover letter, a research statement, a teaching statement, and letters of recommendation. I'll go through these in detail below.

In case you're curious, I posted my original (2002) faculty job application materials online here.

These days, most departments accept the job application online, either via a web form or email. When I applied, only about half of the departments accepted email and I had to send physical copies of my application to the other places.

The first critical component of the job application is your personal web page. I am always amazed at how many faculty applicants fail to maintain an up-to-date web page with their publications, research interests, source code releases, and so forth. Never assume that hiring committees will have your "official" application materials at hand: These days it's much easier to Google someone's name and look at their projects and papers online. For that matter, always post your job application materials prominently on your web page. In case someone is reviewing a set of candidates and can't find your research statement, everything should be linked to your web page so people can find it easily.

The curriculum vitae is probably the easiest part to get right. This should be a detailed summary of your research interests, publications, talks, service work, teaching credentials, and any other factoids that might be of interest to the hiring committee. Under no circumstances should it be a one-page "resume". My 2002-era CV is here as an example. Note how it provides a one-page summary of my research interests and a detailed breakdown of my job experience. The "invited talks" section is provided to give a sense of my experience giving keynotes and lectures at various conferences and universities.

The cover letter is a point of great confusion. First off, it's not always obvious that it's needed, and even when you have a cover letter, not everyone knows what it should say. These days, the cover letter might take the form of the body of the email that you send when submitting your materials. In my experience, the cover letter is a "school specific" statement of why you are applying to this school in particular. It should call specific attention to any potential collaborators at the school you are applying to.

For example, a good cover letter might say something like,
Dear Prof. Zuckerberg,
I am writing to apply for the position of Assistant Professor of Computer Science in your department. My research interests are in the area of computer systems and programming languages, and my thesis topic is "Rooter: A Methodology for the Typical Unification of Access Points and Redundancy." My thesis advisor is Prof. David Culler.
I am excited by the opportunity to teach and do research at University of East Nunavut. My research interests are highly complementary to Profs. Jobs and Ballmer in your department, and I would be particularly interested in collaborating with the Center for Computational Phrenology.
Please find attached my CV, research and teaching statements, and list of references. I look forward to hearing from you. 
You get the idea. It need not be long but it's a good way to customize your application for the specific school, while keeping the rest of your application materials generic.

The research statement is one of the hardest parts of the application to get right. It is intended to serve two purposes: To provide a narrative summary of your research contributions (and especially how they all tie together), and what areas you intend to work on in the future. It's usually about 3-4 pages long and needs to nail what your specific research "angle" is, why the area is important, what your track record is, and what your research vision is going forward. It is not a personal essay like you might have written applying to college or grad school -- If the expression "when I was a child, computers always fascinated me" appears anywhere in your research statement, you're doing it very wrong.

Nobody is going to hold you to working on the specific things you say you want to do for future research directions, but you should articulate a clear vision of what kind of direction you would take when starting a faculty job. This is important. Hiring committees are not hiring you based only on your track record -- they are hiring you based on your potential to be a (potentially) life-long colleague. They want to see that you have an independent and compelling vision for at least the first few years of your faculty job. If the best you can come up with is a couple of papers' worth of extensions to your thesis, you're in trouble. Try to think of a three-to-five year agenda that would get people excited to have you part of the faculty.

The teaching statement is like the research statement, but focuses on teaching. Most grad students have precious little teaching experience beyond a couple of semesters of TA work, so it's kind of hard to say much. Still, do your best. Keep in mind that teaching is a huge part of a faculty job and one of the most important criteria for extending an offer is whether you can teach well. If you have advised any undergraduate researchers or mentored junior grad students, include this in your teaching statement, as mentorship is important too. Finally, be clear on what kinds of courses you would be willing and able to teach. It's not always obvious based on your research background if you could take on, say, the OS or databases course -- make it explicit.

As for letters of recommendation, you usually need three or four. Resist the urge to have more than four rec letters: More is not always better, in case anyone writes anything to give the hiring committee pause. In general it is best if all of your recommendation letters are from well-known professors. Obviously one should be from your thesis advisor. A letter from a top-flight researcher in an industry lab is fine, too, but you should have no more than one of these: It's commonly held that industry folks write fluffy letters and hiring committees care more about the opinion of dyed-in-the-wool academics. One piece of advise I got when applying for faculty jobs was to have one letter from someone not at your home institution, who could comment more broadly (and objectively) on the impact of your research. I was fortunate to get a letter from the great Geoffrey Fox, whom I had met a couple of times and my advisor suggested would be a good "external" letter writer for me. It was kind of strange asking  a near-stranger for a letter like this, but he agreed and I guess it did the trick, since I got interviews pretty much everywhere I applied.

Keep in mind that the job application only gets you an interview, it does not get you the job. The interview is far, far more important than the application materials. It's also important to understand that hiring committees at top schools get many, many hundreds of applications, from all over the world, for a single faculty job opening. So, make sure your packet stands out. A strong publication record is the main thing. Strong letters are second. The research and teaching statement matter much less, so don't stress over them too much. You can't make up for a weak publication record with a brilliant research statement.

Finally, a note on where to apply for jobs. I often see students make the mistake of only applying to the top five or so universities, with the idea that they could only be happy at a place like MIT or Berkeley. This is a huge mistake. First of all, the probability that you're going to get a job at your "top" school is vanishingly small, considering the number of qualified applicants and scarcity of jobs. Second, you might find out (as I did) that schools that look great from a distance don't seem so hot when you're up close and interviewing there. This can cause you to seriously rethink your preferences for both what kind of school you want to be at, where you want to live, and where you see yourself building an academic career.

The converse is also true: You might fall in love with a place you would have never considered seriously before. For example, I knew next to nothing about Harvard before I interviewed there, and never imagined I would end up there -- until I visited, and found that I loved the place and the people. So try to keep an open mind about where you might go. There are lots of great departments out there, lots of great places to live, and many, many factors that count towards your overall happiness and ability to be successful. Apply broadly, include a few "safety schools" in your application list, and then cull the list later if you end up with too many invitations to interview. Most people don't have this problem, so don't be too picky.

Check out Part 1b: How To Get an Interview.


  1. "It should call specific attention to any potential collaborators or..."

    or what? You trailed off at this point, probably a typo. I ask because, other than naming potential collaborators, this is the part of the cover letter that vexes me the most. I read a lot of "faculty application advice" that exhorts the importance of tailoring the cover letter to the school. However, other than naming potential collaborators, I have no idea what to mention that makes one particular school a good fit. This is especially problematic because I work in a small specialized field and many schools will not have anyone who does the same kind of work as I do.

    1. Oops, sorry - fixed the typo.

      The other thing the letter could do is, if you are looking for a specific kind of school (e.g., "I really want to teach at a small liberal arts school") then you should call attention to that.

    2. Note that while the cover letter is a nop at research schools (basically an identification of research area for binning) at teaching schools its really important (in particular to convince them, esp if applying from an R1 Ph.D. program, that you are really serious about this career and don't really want to go to a research program)

  2. Interesting, I would like to know How you felt about Google at that time? of course it is start up and not of uncertain stuffs. but did you basically believe them in general and felt they were technique strong enough?

    1. I didn't take Google seriously at the time because it was "just another search startup", and I had seen several of them fail (Altavista, Inktomi). Back then they weren't doing Gmail, docs, Android, etc. and had not yet become the major attractor for top CS talent that they are today.

  3. Interesting post. As someone who is currently on the job market and more interested in teaching than research, I wonder if anyone has thoughts about similarities and differences between the process Matt describes and the process of applying to teaching-focused departments (i.e. approximately departments without PhD students)? Presumably the emphasis on teaching record and research record are roughly swapped, but I have heard that some small schools put a pretty heavy emphasis on research. Obviously most hiring committees would prefer someone who is strong in both, but realistically I expect there are differences. For example, I hope that a candidate with no actual teaching experience (not just TAing) would be a non-starter at teaching focused schools, just as a candidate with no publications in competitive venues would be a non-starter at R1 schools.

    1. I was our search committee chair at a teaching-focused, (not super prestigious) liberal arts college last year, so I can give you my perspective.

      Cover letter: super important, probably much more so than when applying at research schools. You should convince the search committee you are interested in teaching and that teaching schools aren't just your backup option. You should also try to make it clear why you are interested in that specific institution.

      Teaching statement: Also quite important. Be sure to mention which classes you are most interested in teaching AND which classes you are able to teach. At a small teaching school with just a handful of CS profs, you need to be flexible and willing to teach a wide variety of classes. I don't recommend saying something like "My research area is X and therefore I'm only interested in teaching courses related to X."

      Research statement: not as important, though we looked at these with undergraduate research in mind. I would recommend mentioning ideas for undergrad research projects when you discuss future work.

      Letters of recommendation: I honestly didn't find these super helpful.

      Emphasis on research: our institution requires a couple reasonable publications for tenure and then a couple more for promotion to full, so we mostly looked at the research record to answer the questions "Does this person do legit research?" and "Will they be able to succeed here (get tenure, get good yearly evaluations, etc)?"

      Teaching experience: having teaching experience and including very strong student evaluations in your application is definitely a plus.

      But I'm not sure why you say "For example, I hope that a candidate with no actual teaching experience (not just TAing) would be a non-starter at teaching focused schools." Why do you hope that? Presumably because you have teaching experience and others don't? There are a lot of people who don't have the opportunity to be an instructor before leaving grad school (I didn't).

      Our Dean actually tells us not to focus too much on teaching experience in the searches. We are looking for outstanding teachers more than experienced teachers. For example, if we have 2 applicants, one who has been an instructor 5 times and has pretty average student evaluations, and another who has been a TA numerous times and has glowing student evaluations, then the latter sounds much more promising to me.

      Of course, the actual phone and in-person interviews are what get you the job, not your qualifications. This means that the way you talk about teaching during the phone interview and your teaching demonstration at the in-person interview are by far the most important.

      This is all, again, my own perspective, so your own experience might differ. For example, I'm sure there are teaching schools that won't even look at candidates without major teaching experience.

  4. I happen to have saved a 2002 email from Matt explaining why he didn't want to interview at Compaq WRL. Not sure if I should quote from it without Matt's permission!

    1. Oh dear. I should probably go crawl into a cave first. Jeff, you have my permission to post that email posthumously :-)

  5. As someone who is in the academic market this year, thank you for doing this. Waiting for the remaining pieces..

  6. Thank you for sharing. You really have super strong references. For myself, I have one assistant professor two associates and only one full professor. Two from my departments, two from another two departments of two schools, and one from industry. I either took their classes or worked together. I asked some very well-known professor(s) who are even in my committee, but they are too busy and are reluctant to write letters...

  7. By the way, my publications can be said as very strong (32 journal articles in good journals and I have previous 6-year industry experience and I am graduating in next semester from a major 3rd ranked engineering school in US). But I still don't hold much hope to get into a top schools,as said my advisor and other references are not known at all...

    1. I actually appreciate your post, and look forward to your part II introduction.

  8. Eagerly waiting for part 2 and 3.


  9. look forward to your next part...thank you


Startup Life: Three Months In

I've posted a story to Medium on what it's been like to work at a startup, after years at Google. Check it out here.