Some Useful Tips about Academic Job Interviews Brian Bershad Department of Computer Science, University of Washington Presently at Carnegie-Mellon University Brian.Bershad@CS.CMU.EDU These are the notes from a talk I gave on June 6, 1990 about interviewing for faculty/research positions in computer science. I spent about 2.5 months on the road in 1990 visiting 11 schools and 3 industrial research labs looking for a job. You should consider this note as supplementary to Schwartz's. The comments/suggestions included here are not guaranteed to be anything more than comments and suggestions. Before you go interviewing, you should talk to people directly about what the process is like. These notes will just give you one more perspective on the whole process, what seems to work well, and what doesn't. Above all, interviewing can definitely be a good time. People from all fields of computer science will try very hard to feign interest in what you are working on and will ask you questions that touch on aspects of your work that you would have never thought of on your own. Also, you'll eat very well. Brian Bershad (Brian.Bershad@cs.cmu.edu) - - - ------------ Oh The Places You Will Go Step 1: Getting your name known. You should think of the interviewing process as running from about a year before you will finally end up somewhere to May of that year. Try to get yourself to some high visibility conferences in your area in the 6 months of so before you start to send out letters. Ideally, this means delivering papers. Less than ideally, it means just going to these things and getting introduced to the various people who are in your area the different schools. You should bring copies of your most recent papers (or any if you don't have any that are recent) and be prepared to start answering questions or volunteering information about yourself. Ask the people you chat with if they plan on hiring anyone in your area in the next few years. It doesn't matter what they say, cause they probably don't know and you should apply even if they say they aren't. The point of the question is to make it clear that you are getting out and are interested in coming. Get people's business cards. Don't lose them. Ask them if it would be best to send you stuff to them directly or to the head of hiring (it's always better to do the latter, but again, this is just a leading question). In any case, you should remember with whom you are dealing because you are going to need to hassle them again in the future. Step 2: Where should I apply? Everywhere. Apply to places where you know you will get an interview. Apply to places where you won't get an interview, but might. Apply to places where you might never want to work, but you know there's someone there who you want to meet. Your advisor should be able to tell you what he/she thinks your chances are of getting interviews based on your area, what you've done, and the strength of the letter they will write for you. Push them on this and get an honest answer. You are basically looking for an upper bound. Some areas are going to require 30+ applications, because the there are less jobs in that area and more people coming out and so the competition is harder (in 1990, theory had this problem, but it is guaranteed to change. In the next few years, os and architecture and programming languages look like they are pretty hot. os and arch especially. People in these areas should probably not have to apply to as many places as theory folks. ) Academic Vs. Industry Even if you want to go into academics, it makes a lot of sense to interview at a few industrial r&d labs. here's why: 1. if they really want you, they will show and tell you all kinds of confidential stuff that is lots of fun to hear. 2. if they make you an offer, and even if you don't take, you are in a better position because: a. you can use the offer to hasten offers from universities b. even if you don't take the offer, if they liked you, you will be able to go to them in the future for: 1. consulting 2. matching funds 3. a job if you don't get tenure or hate the university life 4. free hardware The Application: Construct a CV. The important parts of a CV are: Name, Personal, Address, Dissertation Title, Advisor, Prof. experience Awards Publications Research Summary Thesis Research: A brief synopsis of your thesis. General Research Interests: Expand what you say in your thesis and talk about what else you are interested in. It helps to have pubs that back up your interests. Future Plans: Make up something believable and original here. People who will write you letters Make sure you get their permission first. Let them know to all the places that you are applying. They may have suggestions about where else to apply. Recent papers: Hard to say. Some places (Berkeley and CMU) ask for papers explicitly. Others don't care. If you have one or two that you think are really good, you can send them along in the packet. It's unlikely anyone will read them though. The Letters Three kinds of schools: 1. you send them the letters 2. they send out for letters after receiving your packet 3. they expect you to send out for letters after you send in your packet. You need to figure out what kind of school it really is. For example, MIT just sits and waits. Eventually, I found out what was happening: they said that they were waiting on the letters. Depending on how close you are to the people who are writing your letters, you may have to hound them to actually write them. If they don't know you very well, (why are they writing you a letter??), you may want to send them a copy of your CV. The Gestation Period (Number 1) Places get back to you at alarmingly different rates. Since this is your first interaction with the different places, it is worth paying attention to how they handle the application process: a. a good sign is if they send you mail or call you right away acking the receipt of your packet and saying that they will be getting in touch with your shortly. This is unlikely though, given that schools get several hundred applicants. b. a bad sign is if you hear nothing for two months, give them a buzz (this is where it really helps to already know someone on the faculty) and find out that your packet has been sitting on someone's desk for 6 weeks while they are on sabbatical. They will be getting back to you shortly. Eventually, places should start getting back to you. Some will say no. Some will say yes and ask you to give them a good date. Find out if they do one or two day interviews. In general, you should plan on blowing off 1.5+ the number of days in the interview for each interview. Some scheduling tips: Start off light. Take your first two interviews to figure out what you are doing and what the interviewers are doing. Also, use this time to get the bugs out of your talk. Someplace far away is good cause it will give you plenty of time to think about the things you should have been thinking about all along while you are on the plane. Try to do your first interview at a place where you know one of the faculty really well. If they are nice and compassionate, they will understand that you are nervous and will do a lot to make things better. If they invite you to stay at their house, and you are comfortable, you should. Once you've done a couple of lighter ones, then start into the harder interviews. Be careful of scheduling too much in a short period of time. More than 2 interviews/week is a bad idea. More than 2 weeks of two interviews per week is a bad idea. Interviews on either side of a weekend are a bad idea. Two interviews with no intervening time off is a bad idea (I did this at several places and was totally burned out by the middle of the third day). Try not to spend too much time flying back and forth between the coasts. I managed to do all my east coast interviewing on a single trip. This was almost a month long, but there were some advantages: 1. I avoided lots of trips to the airport. 2. I got to spend a fair amount of time visiting friends I have on the east coast. Money: The thing to remember is that it is their's. This means that you should not try to spend too much effort economizing when you book your travel plans. 1. buy tickets with no restrictions. You will on occasion need to change and cancel. It's just a lot easier if you can do it over the phone. 2. if you want to rent a car while you are there, rent a car. While I was on the east coast, I rented cars for almost the entire time I was there. I found that the drives from city to city gave me a lot of time to think (it was basically the only time I was alone), gave me some freedom to visit friends, generally took less time than flying, was a lot less hassle, often cheaper than flying and a lot more fun. Some things to be aware of when renting: will they rent to people under 25? will they allow you to drop off elsewhere? will they let you take the car out of state? Money vs. credit: Of course, although they will pay, they won't pay right away (although a few schools will book and pay for the airline tickets if you ask them. MIT was one of these. CMU was. ) This means that you are going to have to either have a lot of money of your own, or a lot of credit. In my case, I had neither, and ended up borrowing about 8000 dollars to pay for airline tickets. Get an AMEX card. You will end up having to pay for a lot of incidentals while you are traveling (hotels and the like) and it is just easier to know that you are not about to exceed your limit. Also, AMEX likes it if you spend lots of money. Save your receipts for absolutely everything. From the cab to get you to the airport to your beer while waiting for a plane to your espresso on the way to the interview. In general, you won't be paying for lots of things like this, but sometimes you will. You should try to charge stuff to the room if you can. Some places do a direct bill, which means you never see the costs. This can even include long distance phone charges. You might want to be careful here. Send whomever a bill as soon as you get back. Keep really careful track of who owes you what and why. Make copies of the receipts. They will want a cover letter, a total, a simple breakdown, and a social security number. Some places will ask you for your ss#. I got into a habit of always asking them why. Usually, they don't know. Arrival, your Host, and your Confidant. Arriving: If you are flying, then you can get picked up at the airport or drive in a cab. I found it easier to get picked up, but more relaxing to take care of the transport mysel f. If you are going to do it yourself, MAKE absolutely sure you know how t o get from the hotel to the school. Sometimes its easy. Sometimes its n ot. The Host: This person is responsible for arranging the interview and making sure things are going well. You will be corresponding with this person quite a bit before you arrive and he or she will be the one who makes sure that you are taken care of. The Confidant: The person from whom you will get the straight scoop. Often, it will be the host. Sometimes it may be someone on the inside whom you already know. Sometimes it will just be someone you meet while interviewing with whom you get along really well. The Interview Talk The interview talk is extremely important. You can not afford to screw it up. It should be incredibly simple. Motivation (must be extremely clear) Your Thesis Your Contributions Your Work Conclusions Stuff that is optional (I think) is: Related Work Future Work It often comes off sounding hokey when people include this. You have only a limited amount of time in your talk with which to get your demonstrated ideas across. Having a related work section just shows you can do legwork -- that's to be assumed. A future work section can be ok, but can also be dangerous unless you have really thought hard about the direction in which you expect things to be going. Key goals: CONTENT: It should be under 50 minutes. Most places will give you an hour. You will get questions. The entire thing should be understandable by everyone. This means that you don't need to give too many details. When it is necessary to get detailed, return frequently to the high level so that if you lose people, you don't lose them for too long. (There is another school of thought here that says that you should get progressively more detailed until only one other person understands what you are saying and then conclude. This is a terrible idea.) There should be no code. There should be no numbers. Graphs are good. they should be simple. Histograms convey the essence of numbers without the decimal point -- use them. Your conclusions should clearly sum up what you are doing. Your talk is to convince the people not in your area that you are a good person. The people in your area should be able to determine this from your interview, from your thesis, and from your publications. If you are a systems person, these are the kinds of things that makes non systems people feel good: 1. a good systems/performance talk is like a proof where you have demonstrated a new lower bounds by taking a new approach. 2. theorists have a habit of thinking that systems work is engineering. Show them that your work is not just engineering, but demonstrates insight and original thought. If you are a theory person, show that there are some practical aspects to the problem you are working on. Perhaps start off with a real world example motivating your existence. PRESENTATION: In some respects, presentation can be as important than content. If you have a brilliant thesis, but your title slide is spelled wrong, you are hosed. Your slides should be very simple. No overlays. They always get misaligned. Your basic talk should be nearly memorized. You will have 30 to 50 slides and you should be able to sort them without having numbers on them (visible numbers are a bad idea) You should be able to give the talk in your sleep. Here are some things that I did that really helped: I had my talk memorized. I didn't give it as though it sounded memorized (except on one occasion where I was so burned out that if I hadn't of had it memorized I wouldn't have had anything to say) For each slide, I had the original sitting on top in the stack. Each original had special "key point" annotations that I wanted to be sure to say. I made sure that I had a big space around the slide projector. Big enough for me to have my old slides, the sheet of paper for my current slide, and the sheet of paper for my next slide all visible. This way, I always knew where I was, where I was going and was sure to not forget anything. I made sure that nobody was able to sit directly in front of me or to my side so that they could look at my slides before I got to them. I found this incredibly annoying. I gave my talk in front of people here at least twice. I gave my talk in front of an empty room and a video camera. This is critical for a couple of reasons: it forces you to run through the whole thing when you really don't feel like it. You will experience this feeling on the road. it forces you to give a talk to a roomful of empty faces. This will happen at least once. I was able to review the tape, watching for excessive motion, side comments, slide slapping, miscues, words that sound funny, critical sections that you screwed up, etc... VERY IMPORTANT: When scheduling my interviews I made it absolutely clear that I needed the 30 minutes JUST BEFORE the talk to be left alone. No interviews. Nothing. This was my time to collect self and get ready. What I tried to do was get put into an empty room and run through the talk at double speed. You may find that you get to places and that this has not happened. FORCE THE ISSUE. Even if it means cancelling an interview with someone. THEY WILL UNDERSTAND and if they don't, then you probably don't want to work there. If you don't do this, and try to enter a room cold and give a talk that you have memorized and have given a dozen times already, it will be a disaster. I promise. PROBLEMS DURING THE TALK The Designated Jerk Some places will have a self-appointed designated jerk. This person is easy to recognize. He will start off at the beginning of your talk asking you to define stuff that is either obvious or unimportant. He will then proceed to interrupt you every five minutes trying to make it clear that your work is silly and pointless compared to his. If you are lucky, the audience will include the designated jerk detector and destructor. This person will take responsibility for recognizing the jerk and telling them to shut up. When to Be Rude: When the designated jerk has overstepped his bounds, or is heckling, or is asking questions just for the sake of asking questions, the best way I found to be rude without being rude was to start giving one or two word answers to questions. The danger here is that some people will not sense that you are responding to a jerk and will think that you are a jerk. Don't spend too much time worrying about this. The Designated Idiot Unfortunately, there will sometimes be audiences which include people who will never be able to understand what you have done, no matter how you explain it to them. These people are hopeless and all you can do is be polite. Try to answer their question as quickly as possibly and hope that everyone else recognizes what you are up against. Under no circumstances should you be rude. Props: You should have a pen, some water (plenty) and whatever props you happen to need OUT and VISIBLE before the talk begins. I used a penny in my talk, and when I was giving one talk, the penny was nowhere to be found when I needed it and ended up having to borrow change from the audience. It worked out ok, but could have been a disaster. Questions: Anticipate the hard questions and have good answers for them. Try not to answer questions in a tone/style significantly different than your delivery style, otherwise people will think that you are scripted heavily. In the best of circumstances, questions will punctuate the entire talk. Sometimes, people will be absolutely dead until the end. It is more of a style thing and you should not use it to judge how well you are coming across. At Cornell, for instance, nobody said anything and the entire audience was on the left hand side of the room. It was not until the end, when there were 20 minutes of questions, did I realize that everyone was paying really close attention... they were just waiting for me to finish. Recovering from Disaster: Nothing is as disasterous as it might seem to you while you are giving your talk. You will forget stuff. Nobody knows but you. You will say things unclearly. People will know, but will think that you are just being unclear and that they would understand if they were listening more closely. People will only remember 3 things from your talk: 1. were you having fun. 2. did they learn something new 3. did you have any contributions. They will not remember anything else. You will, but you can't let it phase you. The Bad Talk You will inevitably give a bad talk. The bigger the place, the less likely it is that there will be a lot of people outside your area there. In my worst talk, the rule of thumb from above which says make the theory people happy ended up as a bad rule. I ended up disappointing most of the people in my talk because it was at all much too high of a level. They wanted implementation details, they wanted to know how it worked. Further, I was tired (bad scheduling), did not get my 30 minutes before (bad timing), and there was a very small turnout. Small turnouts are quite hard to deal with. Not too many faces and it is discouraging when they leave early. After The Talk If you are lucky, then you will have some winddown time right after the talk. You will definitely need it. If you are really UNLUCKY, then you will have interviews right after the talk. For me, I was always a zombie right afterwards. I was tired, on high energy coming down, and not able to think in one-on-one mode. If it happens, excuse yourself to go to the bathroom for a few minutes and calm down. (The bathroom is also a good place to go between 30 minute interviews; otherwise you will never have any time alone during the interview days.) The Interviews: You are going to meet a lot of people. The worst places will schedule 8 to 10 people per day for 30 minutes each. Other schedules are 45 minutes and an hour. There is no good rule about which is best. If you are having a good time in a 30 minute talk, then it sucks when it ends. The opposite is true of the hour long interview. Be prepared with lots of canned questions for when the interview doesn't go well. You should figure that it is the interviewer's responsibility to make you feel comfortable and to run the interview and ask questions. When things drag, it is their fault (this isn't always the case but you will feel better if you think this way). Some good questions: Tell me about your work? Are you happy here? How does the dept. leadership work? What do your days look like? IS this a good place to be an asst. prof. Hard questions will tend to get answered honestly by people who are really happy or really unhappy. They will tell you. People who are in the middle won't feel comfortable or won't want to answer them. In this case you shouldn't even bother asking. Try to pick up on points about the interviewer's life from clues in their office. For me, for example, I would look for squash rackets, bikes, motorcycles or kids. Each of these were things that I knew enough about to have a good social conversation about. It helps to start off on a non technical note, so that you can develop some personal rapport. Of course, if you try to develop personal rapport with 12 new people a day, you will undoubtedly come up against some duds. Comp Sci has more than its fair share of personality zeros (I am not sure why), so when it happens to you, try to stay on a non-personal topic. If the person is outside of your area, try to find out what they are teaching and get them to discuss their experiences there. You should be able to respond or give feedback to what is happening in the undergraduate curriculum, for example. Some people you will meet during your interviews: The Lazy Interviewer: Your trip has been scheduled for weeks and the people who are signed up to see you have had a long time to figure out what you do. Each person you see will have a copy of your packet which they should try to read before you show up. Often, you will do interviews without having given your talk yet. The really lazy ones will not bother to read your packet and will want you to explain to them what you do and what your talk is about. I often found myself reciting the whole intro and motivation and conclusions of my talk in the one on ones with people. Professor Deadwood: You will know him when you meet him. He will tell you stories about when they were first bringing up cobol on the ibm 360 and how things today really haven't changed much. Just nod politely, try to bring him up to date and hope that this is a 30 minute interview. Dr. UnbelievablyBad Occasionally, you will run into someone who is unbelievably bad. They just won't have a clue about what is going on or the way things work. Chances are, though, that if you recognize them as being really bad, then the rest of the faculty does too. If they are not in your area, it's nothing to worry about. If, though, they are in your area, you should be really concerned because it indicates something is seriously wrong somewhere. If the guy is tenured, you have to ask why? If he isn't, you have to ask when he is leaving. If the place you are going to is interested in hiring you, they should be willing to talk candidly about their "problem children." Dr. GoodComputerScience: This will be a Great Interview. You will remember it forever. Perspective, history, or current state of the art. The Why Am I Here Interview: Personnel Want to know $$ information. Don't say anything other than it is too premature to be discussing these matters. The Dean of the School Can you walk and chew gum at the same time. Be prepared for layman questions. This interview is designed to ensure that the dept. is not trying to snooker the university by padding the faculty with vegetables. The System Manager What kind of hardware do you need. Treat this as a personnel interview. The Visit With The Chair: At each place, you will have a visit with the dept. head. It is his job to answer any questions you have about protocols, teaching load, general requirements for the first year, salary, PYI's etc. (remember, everything is negotiable, but it is the chair's job to tell you where the negotiations would start). You should also tell the chair about any constraints you might have: 1. two body problem 2. late thesis 3. need to hear a decision by,.... The chair will probably ask you where else you are interviewing. You should answer this. You might also ask who else the dept. is interviewing, just to see what kind of a response you get. You should ask in what area is the dept. growing, what are the goals, how things work, does the chair have any real power (ie. control over teaching assignments, discretionary money, control over promotion, etc.) Find out about housing deals. Some places have them. Some don't. Socializing: You will eat a lot. Lunch will be the faculty club. Eat light. Be careful. Best to not drink anything. Dinner will probably be a very good restaurant. They might ask you where you want to go. Defer, because they will know better places then you. Don't even say what kind of food. Dinners can be great or they can be hell, depending on the chemistry of the people who are there. If the people all know one another and work together, and are reasonably tight knit, then dinner can be super. If there are more than a few people, they should be trying to draw you into the conversation to see how you act in a social setting. At some of places , dinner was with too many people who barely knew one another at a noisy dinner which was pretty much dominated by one individual and that wasn't me. I found this to be really depressing and I was down on myself for a couple of days. It's perfectly ok to not be able to be a full participant at dinner with 4 people after 10 hours of interviewing. There will probably be a lot of wine ordered. If you can drink well, then you should have some. If you can't (i.e, it gives you a headache, or makes you sleepy, or horny, or ...) you should definitely abstain. It can be fun to watch the faculty get stupid while you are doing well. On the other hand, it can be fun to get stupid with the faculty. The Dessert Party: Or Socialized Torture No matter what people around here think, the dessert party is really the worst possible thing that can happen. After 14 hours of talking, interviewing, and eating, you all of a sudden find yourself thrust into a room of faculty and graduate students with lots of cookies and beer and more people who want to ask you questions. While it's an incredibly nice gesture on the part of the people who are giving and attending the party, it is also very wearing. Just hold your breath and try to get out of there before 10. Day 2: It starts all over. The morning of the second day for 2 day interviews is the worst. I was still tired, I felt as though all of my questions had been answered on the previous day, was tired of hearing the same questions, and wanted to go home. Usually, the second day would start out real slow. It helped to just come right out and say this. In general, meta conversation about the whole process worked to make things a bit more relaxed. People who had just been through the process will tell you what they did, and you can complain to them about how hard it is, but how it really is a lot of fun and... The Offer/Rejection phase: Here's how hiring works: Each place has so many slots. The faculty decides on its top choices and hopes that there are more slots than choices. Some places, like MIT and Cornell, in effect have as many slots as they need. Others, like Berkeley, run really slot limited. Someone will contact you and find out if you are interested in getting an offer (this is for places that have too few slots). Here are reasons to say yes: 1. you have no other offers. 2. you might want to work there. 3. you really want to work there. 4. you might want to work there and think you can get a startup negotiated which you can use as baseline to guage other places. If none of these are true, then say NO before they make the offer. You may hurt someone's feelings, but in the long run it makes life for everyone easier. It costs to make an offer that won't get taken. It's free to not make such an offer. If you say YES, get the offer in writing. If the offer has a time limit on it, either try to get a long limit, or delay the official offer. Both are possible. If they really want you, they will defer the limit as much as you need. The Negotiation: Once you have said Yes, one of three things will happen: 1. you will be presented with a "standard" startup package. This probably includes: 1. a workstation 2. light first year teaching 3. some conference/travel money 4. some discretionary money 5. 45-50k for nine months 6. guaranteed summer support for one year. 2. you will be asked to come up with an initial startup. It should be a superset of the standard. You can ask for all kinds of stuff here: space, extra $$ for students, hardware for your home, lab equipment, lighter teaching, more discretionary money, the works... Try not to go overboard. 3. you will be told that your research needs will be taken care of for the first xx months. Case closed. Few places make such offers. For these places, there seemed to be no real need to negotiate a hard startup. Basically, everything gets taken care of and whatever you need is handled. This happens only at really rich schools. If they really really want you, they will be willing to put forward "whatever it takes" to get you there. We can be talking a few hundred k if you can really show that you need it (this usually goes to lab space and equipment). The way startup gets paid for is the chair goes to the dean and says "we need XX dollars to get this person here. See, here's the startup proposal they gave us." The dean says "ok, but that's an awful lot, how about if I kick in xx percent" and the dept. kicks in the rest. " So, its really a chance for you and the dept. to get some free money from the dean. Again, don't be afraid to ask for whatever you need and can justify. The Phone Calls: Once a place has made you an offer, they have decided that they want you on their faculty. This mean that they will contact you ceaselessly until you make some kind of a decision. Figure 3 guys at each place, each calling you twice a week. Figure you get 4 offers. That's 24 calls per week, each wanting to know if there is anything else you need to know about whatever that would help you to make a decision to come to their school. It gets really hard. Don't be afraid to say that you are just trying to think things out and would really much rather be left alone and if you do have any questions, you will call them. Again, they should understand and if they don't it's indicative of a bigger problem. Think of how strange this would be if the sitation were reversed. You have interviewed there, they have not made an offer yet, but you are calling them 3 times a week asking them if there is anything else they need to know about you. Completely inequitable! A Second Look: If you are thinking seriously about a place, then you should by all means go back for a second visit. Most places will suggest this, but if they don't suggest it to them. If you have a spouse or someone with whom you will continue living when you move, take them along to see the city and to watch your reaction to what happens. This is your chance to make a low key, low pressure visit back and get a sense for how things might be if you were really there. If they don't offer, ask the schools to pay for your return visit. It's a low cost investment in the future. You might spend 3 days in a place, stop in to see some of the people in your area on one morning of one day, and spend the rest of the time just hanging out getting a feel for the place. Making a Decision: I will not try to tell you what should be the criterion in making a decision. It's a very personal thing, and depends on a huge number of factors. In other words, I am copping out here. Saying Yes: Saying yes is easy. Call up the chair. Say yes, that you will be glad to accept and are looking forward to coming. News travels fast, so you had better start the no phone calls right away. Saying No: Saying no is very hard. Reactions tend to be as though you are rejecting the individuals at the places personally (which you may well be). Be prepared with a list of reasons as to why you are saying No. Sometimes I was asked just out of curiousity, but sometimes it was because the dept. wanted to know what they could do better in the future to improve the chance of people saying yes. Be honest here, because people can tell when you are bullshitting about this kind of stuff. A few people will try to get you to change your mind. That's why it's important to say Yes first to someone else. You should be glad that you did not go to one of these "arm twist" places. In the worst case, they will tell you what a bad decision you have made and they are really surprised. This is the worst, since you probably s till have doubts about your decision and it doesn't take much to set them of f. Try to ignore these people. Usually though, the people will be very professional about things and understand. Besides, if they really wanted you, they have to know that they should not cut their ties by setting you off, since you one day may feel like moving again. It is a very small world. Finishing Up: Finish your thesis before you go and then take a vacation.