Advice for Landing a Faculty Position

• Find a home: You need to have a "home community", or sub-discipline. This is typically defined by the kinds of research problems you pursue, or the set of techniques you use to solve the problems, or both. More concretely, your home community is defined by a set of conference venues (e.g., SIGCOMM/NSDI/CoNext, or FOCS/STOC/SODA, or whatever). You'll want the bulk of your publications to appear at a set of related venues (and preferably the more competitive and visible venues), so your work reaches your home community and you are visible within that home community. Later, when you apply for a faculty position, the faculty in this "home community" can push for you to be hired, and view you as filling "their slot" if hiring is done by area. Even if you do interdisciplinary work, junior researchers should "pick a side" and focus on one community for building their publication record, professional visibility, and so on. For example, if you apply theoretical cryptography to security problems in computer networks, you should decide if you are a networking person (and publish primarily at SIGCOMM, CoNext, INFOCOM, etc.), a security person (and publish primarily at CCS, S&P, USENIX Security, etc.), or a cryptography person. As you gain more professional stature, and start advising students of your own, you might choose to straddle two related communities (e.g., networking and security, or networking and distributed systems) and their associated conferences.
• Explain yourself: You should have a sense of how to "explain yourself" to others. What is your research "taste" -- what kinds of technical problems excite you, and what approach do you take to solving them? This should be more than merely describing yourself -- this should be something you deeply believe is a valuable set of technical problems and solution techniques, so you can convey to others why they should want a person like that in their department. This explanation should appear at the beginning of your Research Statement when you apply for faculty positions. Typically, the first couple of paragraphs of the research statement tells a story about what kind of person you are, and why. My students and I spend a lot of energy on brainstorming the "story" they tell in these statements, not only to have a strong research statement, but to have a good "elevator pitch" to describe themselves. Figuring out your "research taste" (which may very well differ somewhat from your advisor's research taste) is an important "take-away" from graduate school or a post-doc.
• Think big: You'll want to have a longer-term vision. More than a bunch of top papers at top venues, you should have some research theme that makes your papers more than the sum of their parts. Having a strong publication record is, of course, valuable, but flitting from one topic to another has its costs in terms of the depth of your work, and (ultimately) the strength of your job talk. I have seen various PhD candidates with top publication records under-perform on the job market because they haven't dug deeply into some important and difficult problem, and truly nailed the solution. Many times a PhD thesis is just a "staple job", where the story linking the chapters is somewhat of an after-thought. While this is not necessarily a problem, you do want at least some significant portion of your work to build toward something bigger than any one paper. That may be a deeper theoretical understanding of some area, or a large system or language you are building. Whatever it is, you should have some Big Thing under your belt, and more Big Things in mind for the future. And, you should make other people want to follow your vision.
• Talk well: You'll want to get adept at giving talks. Many people do great research work, but do not present it well. Or, they present the work clearly, but fail to convey the broad importance of the problem or the novel contribution in the solution. Giving good talks is very hard. It is worth spending a lot of time preparing talks, getting feedback, revising the talk, and so on. For the job talk, it is especially important to give practice talks to non-specialists. Many faculty candidates completely snow the audience, giving a talk that is great for people in their own field, but unintelligible for everyone else. Or, they assume that the audience will appreciate the importance of the topic, failing to realize that only their peers in the same discipline already see the research problem as important. Also, the audience for your job talk consists of a lot of busy, self-centered people, and you need to get them to pay attention---to understand that the problems you pursue, and the way you pursue them, are important, novel, creative, surprising, and all around awesome. You want them to understand that they need someone like you in their department.
• Listen well: You'll also want to attend faculty-candidate talks and distinguished lectures in the department, including talks outside of your research area. Watch how these researchers motivate and explain their research, and see whether they are effective at conveying the excitement of their research area and the depth of their research contributions to others. And ask other colleagues (especially the faculty) in the department what they thought of the talk. You might be surprised to hear they had a very different impression about whether the talk was effective or not.
• Observe others: Look for a "short list" of people you respect in your field, and try to discern the pattern behind what they do. Where do they publish, what kinds of problems do they pursue, what is the "method behind their madness", what is their "research taste," and what kinds of professional activities do they get involved in. These people may be your advisors or mentors, but they do not have to be. Ask yourself why some particular paper is highly visible and widely cited. Was it the first paper that identified a new and exciting problem? Did it have a surprising answer to a long-unresolved question? Was it presented especially well? Meet these people at conferences, and ask them questions about how they pick research problems and how they approach solving them. Most people are delighted when people ask them "meta questions" like this, but it happens far too rarely. On the flip side, consider some paper that you really admire that is not widely read and cited, and ask yourself why. What is the reason that this paper is overlooked and under-appreciated, and what could the authors have done to prevent that? Was the problem too narrow or not especially important? Was the solution solid, but not general enough to apply in other contexts? Was the paper published in a less visible venue? Did the work fail to carry the ideas far enough? Was the paper unnecessarily hard to read?