So You Want To Graduate This Year: Tips for Your Final Year of Graduate School

Job Applications

What jobs are out there?

There are lots of jobs out there, so don't let anyone else tell you otherwise. In addition to the post-doctoral positions I'm about to list, there are also many opportunities outside of the standard track in companies, finance, education, and more. You will get a job, so don't worry about starving on the street!

There are three main types of post-docs positions to which you can apply (in descending order of prestige):

  • National fellowships (e.g. NSF/NASA fellowships that you can take anywhere)
  • Institutional fellowships (e.g. Princeton Spitzer fellowship, Berkeley Miller Fellowship, etc.)
  • Postdoctoral position with a specific task and sponsor

The national fellowships are the most competitive because you are granted freedom to work on whatever you want, and you are not tied to working with a single faculty member (or even tied to an institution). Institutional fellowships are also quite competitive (in some cases *more* competitive), as again you are not tied to working with a single faculty member and you have freedom to work on your own research goals. Lastly, there is a large variation in postdoctoral positions of the third kind, but invariably you are hired by a specific person at a specific institution; however, you may have a specific task to complete, or you may have a great deal of freedom to perform your own research depending on the terms of the job.

When to start?

The deadlines for job applications are spread throughout the year, although the heaviest density are from October to February. Cameron recommends for students who think they will be graduating in one calendar year to begin working on applications in July the year before. Notable early application deadlines are:

Because of its early deadline, the International NSF is oftentimes overlooked, but it is a good position (particularly if you want to work abroad) (This seems to have been discontinued as of 2018). The AAPF is probably the most difficult application out there, as it requires a ten-page research plan, in addition to about 10 auxiliary documents. The disadvantage to this is that it takes a lot of time (Cameron's took about 100 hours), but the benefit is that once you have completed it, you have done about 80% of the effort needed for all of your remaining applications.

Taka recommends starting to put together research proposal at least a month before your first deadline (more, if you want feedback from advisors, etc.). The first big one that people usually apply for is the NSF in October. I'd start thinking about proposals during the summer. David had all of his letter writers immediately ask to see a copy, so you may want to have a draft when you ask - they can comment on its feasibility and value in their letters as well as give a slightly more distant critique than your thesis advisor.

Have your CV ready to go. Take a look at some postdoc and grad student CVs online to see what they have on there, how they format it, etc. Talk to your peers and some of the past graduates to see their CVs (and tex files). Look online at other astronomer's websites and steal their CVs to see what you like best.

Employers will want either a research proposal or a statement of research interests. Most likely, you will have to prepare one of each. A research statement is typically three pages, and it describes the research you have already performed in your time as a graduate student, as well as setting the stage for the next few projects that you wish to tackle. It would be wise to have a baseline research statement, which you modify slightly for each position to highlight your interests/background in the area to which you are applying (and including references to the work of the people with whom you will be working). The job register now (2017) has postdoctoral application guidelines.

Also, it is very hard to get any research done while preparing proposals and other application material.

Where to apply?

The AAS Job Register is a good first place to look for jobs. They update the 1st of every month as submitted to the register, pending payment (as of fall 2017). Archived jobs are not listed or searchable but you can browse them by month to get an idea of what might appear when during the cycle. The Rumor Mill (see below) is also a good way to find out what recurring jobs might be available.

ZH recommended that Taka apply to every job that is preferable to quitting astronomy. I echo that advice. It sounds like a huge pain, but it's not really. Once you prepare and submit ~10 applications, preparing and sending another won't take more than ~20 min. I think 20+ applications is normal. I applied to close to 50, and know others who applied to more.

I found that even jobs I wasn't seriously considering initially can make very convincing and alluring job offers and sales pitches, so keep an open mind.

Be aware of the odds. Popular postdoc opportunities will have over 200 applications for a handful of positions. The average success rate is on the order of a few per cent across the board, so unless you're convinced that you're a truly exceptional candidate, be prepared to send many applications and receive many rejections. The intent here is to be realistic, not to sound pessimistic; Columbia astronomy PhDs? tend to do well. You can see a record of where recent grads went on our website.

Cameron has heard of students applying from as few as 2-3 positions, and as many as almost 100 positions. Realistically, once you get all of your application materials in order, it isn't *that* much additional effort to apply to more positions, so I found the difference between applying to 5 positions and 30 positions to be only about 10-15 hours of additional work, but your mileage may vary. I would recommend only applying to jobs that you could actually see yourself taking if offered, but for some that might be 5 and for some it might be 50. Cameron applied to about 30.

Staying organized

When applying to so many jobs with deadlines all over the calendar, the people writing letters for you (usually 3) might have trouble keeping track of everything.

Taka tried a spreadsheet in Google Docs, with columns for Job, deadline, web link, letter submission method, etc. and boxes where references could enter initials upon submitting their letters. This worked great; two of my letter writers used and liked it (one was super-organized and didn't need it). Cameron concurs. David did the same and added an email to writers every 2 weeks with a short summary which jobs needed letters and how to submit them; these seemed to be well received.

On writing a research proposal

(This section is based on advice I got from professors, feedback I got from the NSF reviewers on my proposal, as well as after-the-fact evaluation of which versions of my postdoc research proposals did well, which ones did not, and why.)

Make sure the proposal has a central focus; it should not be a laundry list of 6 projects you could do (advice source: Greg Bryan). At the same time, it should not be too narrow; that is, it should not be the subject of one or two future papers, but rather a topic on which you can perform influential research for the next 2-3 years.

Don't be afraid of stating the obvious. You should assume that the reader will not be an expert on the subsubfield. (It's rather likely that the reader will have a shared expertise on the same broader subject as your proposed research, but be unaware of some scientific motivations/findings specific to your topic.)

If you are proposing to do "more detailed/extensive" add-ons to past research by yourself or others, be sure to spell out the specific benefits of the added detail/depth. If it is theoretical work investigating "the effects of previously un-/underconsidered mechanisms", make a strong case that the added theoretical complexity will plausibly be useful in understanding observations. (This was a very helpful piece of criticism from one of the reviewers of my NSF proposal.) Most of the proposals people read will be work no one else has done; you want to highlight why your ideas should be prioritized over everyone else's.

Be able to concisely state your qualifications for the research. You don't want the proposal to be about you, but you do want to convince the reader that you have all of the skills, tools and experience to do what you're proposing. This is especially true if your past work is not obviously related to the proposed research. (Both of my NSF reviewers very clearly noted that I am qualified to do the proposed research.)

If you are proposing to do outreach or mass-dissemination types of activities for the NSF "broader impact" component, present a detailed and feasible plan for evaluating how successful it is. This is advice I got from Marcel Agueros, and I tried to incorporate it. Both of my NSF reviewers highly praised my broader impact idea, but expressed concern about evaluating its successes. From the NSF perspective, they have to be able to justify to Congress that the money was well-spent, and have objective information to back up this claim. (This may seem obvious, but it's something grant-granters really think about. I heard from the director of the NSF EAPSI fellowship on this topic.) Similarly, if your proposal relies on getting help from others (maybe you know artists or computer programmers to help with the broader impact component), convince the reader that the arrangements are concretely in place.

If you make a website, make it presentable

There's a very good chance that someone will Google your name if they are considering you for a position, and/or if they see your name on the rumor mill (see below). If you have a professional/personal website, make it respectable and so that people can see info about your research and access an up-to-date CV. These websites can be time consuming to create (and make it high into google search queries for your name), so budget time early for this (i.e. july, august). Your Columbia astro web page will certainly come up too, so you may want to add a link to your ADS and CV there as well.

Also be mindful of what might come up when people Google your name. If you have blogs or Twitter accounts where you make NSFW remarks, you might want to consider hiding those. Make sure you understand your privacy settings on Facebook. I cringe sometimes at the things some astronomers say on their public Twitter accounts.

It's a small field, and people talk. We're the first astronomers to live in an era where young professors are active gossipers on Facebook. I can name several instances of professors complaining about their grad students on FB, and of students and postdocs complaining about their advisors/supervisors; often, it's not too hard to know (e.g. by looking at coauthor lists, gender pronouns, directories) who's being complained about, even if full names aren't given. I had a professor ask me (a grad student!) what I knew about someone they were considering for a postdoc position. I heard from a research group in Europe that the possibility of my joining them as a postdoc was a frequent topic of conversation.

Okay, I've applied. What's the timeline?

The AAS has a recommendation that astronomy postdoc positions do not require candidates to accept or decline the offer until Feb.15. (Note that many physics and European institutions do not abide by this.)

The major fellowships are usually announced ~2 weeks after the winter AAS meeting, so in the last half of January. Most employers make shortlists and interviews in late January or early February, but many do so before the AAS meeting and sometimes as early as Thanksgiving. It's common to receive offers right before or several days after Feb.15, as people decline offers and institutions scramble to contact others on the shortlist.

Once you have an offer, are not 100% sure about accepting it, and are waiting to hear from other places, don't be shy about emailing these places as it is common practice. It doesn't hurt your application status to inquire politely about it. I've heard professors speculate that an inquiry, accompanied by information that you have another offer, could actually help your application.

There is a second hiring wave after Feb.15, with applications due throughout the spring. David found that Spring 2018 was sparse, especially for jobs based in the US.

You will typically be given some sort of interview before being given an offer. Cameron received requests for several skype interviews, but also a few in-person interviews where the host paid for the visit and the job talk. Most interviews are pretty laid back. Be prepared to talk in depth about your research (easy), about tie-ins to your host's research (more challenging), and tie-ins to other faculty members in the department with whom you might want to work (also challenging). It means doing your homework beforehand, and reviewing all of your application materials to that position. Have an idea of when you would like to start the job (some might want you to come early/late, but typical start dates are September 1st). If you are giving a job talk, these are usually about 30-40 minutes long and part of a colloquium or group meeting, so keep it with a broad introduction (as there might be non-specialists present).

To rumor mill or not to rumor mill?

If you don't know already, there is a website called the Astrophysics Rumor mill. It is currently hosted by AstroBetter. It is a place where people (anonymously) post and edit job openings, short-list status, offers, and who's headed where. There are several advantages to this page:

(i) It is a good place to find out if places have set dates for search committee meetings, made shortlists, or offered positions. This can be discouraging information, but it is very practical. Most places take a looong time to tell you if you didn't make their shortlist and are effectively rejected (I've received several in March). You don't want to wait forever for a place to get back to you. It's a giant, chaotic game of musical chairs, and the rumor mill allows you to play the game with one eye open, instead of playing it blindfolded.

(ii) It can be good self-promotion. People from all over the world look at this page, Google the names, look up publications, etc. When I landed on a couple of shortlists, I had an astronomer in Japan (with whom I was discussing collaborations but didn't know personally) congratulate me and tell me that many of his colleagues overseas (some of them on selection committees) have been gossiping about me. While one could raise ethical questions about employers peeking at an anonymous rumor mill to scope candidates and the job market, the bottom line is that some of them do. It is not unheard of for an institution to sweeten an offer once they learn that they have competition for a candidate's services (this has happened at least twice in the last 2 years to candidates from Columbia).

The main disadvantages of the rumor mill are the possible stress of watching the job market evolve, and the fact that occasionally some trolls will post inaccurate information. (Don't pay attention to the ~5 people whose names pop up next to every position, because they can't fill all the positions.) The page isn't for everyone, but Taka suggests it as a potentially useful resource. Taka's personal recommendation is to not pay much attention to it until you are shortlisted for a position.

Cameron found it to be a good resource when not used too heavily. He added his name on offers, so that people would notice/google him more frequently, and it seemed to work pretty well at generating further interest.

--Taka (with additions by Cameron [and David])

Dissertation Completion and Submission

Our department is great in terms of its requirements for the dissertation. In general, you can use your first-author publications verbatim as individual chapters of your thesis. In Cameron's case, he used his two first-author papers and his unpublished first-year project as his three main content chapters. You must also write an introduction and a conclusion to your thesis, which can be somewhat time-consuming, but it is really up to you about how detailed you want to make it. Typical introduction chapter lengths are 10-20 pages (manuscript latex style), although Taka's was around 50 pages. Conclusion chapters are pretty short (e.g. 4-10 pages) as you're not really summarizing your results, as making broad statements about what you learned and what the direction of future work will be on the subject. Talk to your advisor about what you want to include in the thesis, but in general you want all of the chapters to be focused on a single theme.

Graduates typically defend somewhere between April and August. As long as you defend by the end of August, you can walk in that year's May Commencement ceremony, if you're into that whole thing. Otherwise you can postpone it to the following year.

You and your advisor should discuss who you want to be on your committee, but it is typically 5 faculty members 3-4 internal, 1-2 external.

Around March, you will have to sign some forms with Ayoune to declare your intention to defend. No big deal, but you need a title for your thesis (which can later be modified). Work with your advisor to pick a date/time for your defense. This is usually pretty challenging to find a time when all of your committee members can meet (and may decide who can be a committee member), but be flexible. You will be expected to submit your completed thesis to each of your committee members two weeks before the defense date, in order to give them time to read it and critique it.

Once you have completed writing the thesis and distributed it, spend a little time relaxing. For the defense, you will need to put together a short presentation that would take you 20-30 minutes to get through if you were not interrupted. Make it at the level of your committee (i.e. if you have one outsider who is not really familiar with your research, then make it at the level of that person, but not any lower). In general, I think this was the "highest" level talk I've ever given because mostly I speak to a mixed audience and these were all specialists in my field. Make sure the slides have the important results, and go in order through the chapters of your thesis (but you don't have to have EVERYTHING, just the main stuff). For example, I had a slide of *every* plot from my thesis, but many of these slides I included at the very very end (after the ending slide), so that I could reference them IF anyone asked questions about them, but I didn't talk about some of these in the main part of my presentation (they were just there as reference). This actually was useful, as I was asked a question about one of the plots that I didn't present, and I had it in the back, so it worked out.

I spent the day before the defense putting this together, so don't spend tons of time working on this, just put it together in a day, and then practice once or twice.

The defense itself can last anywhere from 1.5 to 3 hours, depending on your committee. All you do is do your presentation and get interrupted a bunch throughout as they ask questions. When you reach the end, they may ask you some "general" questions, but I think this is rare within our department. The only "general" questions I got were from Jerry Ostriker, asking things that I didn't actually cover in my thesis, and he had me do a derivation on the board. But it was fine. I had actually spent a couple of days reading up on some of the derivations related to my work so I was prepared. But again, it doesn't happen to everyone. If you *want* to prepare on that, do some reading up. Since a lot of my work is on prescriptions for star formation and feedback, I read up on how to do the Jean's Mass derivation, the freefall time calculation, the virial theorem, the total amount of potential energy held in a uniform density sphere, etc. I basically read through a bunch of the related chapters in Carroll and Ostlie, some Binney and Tremaine, some Dopita and Sutherland, etc., to put everything in perspective. Ultimately, I was concerned with knowing better some of the details of the assumptions I was making in my research, so from that perspective, it's useful to know even without the defense. So you might work on that if you feel like it.

I'd also recommend looking at the relevant work by your committee members in your field of research. It was good to read a couple of their papers relating to my work, so I knew where they were all coming from. Of course, one cannot read every paper by every member, but just the relevant highlights. Ask your advisor for what works might be most appropriate from a few members.

Most of the questions will be on *your* work, but a few on the assumptions you make and the other papers out there that you're building on. I'd say as long as you can answer all the questions about what *you* have done, you're fine. You're allowed a little bit of leeway on not knowing the literature in its entirety.

At the end of your defense, you'll be asked to leave the room while your committee decides your fate. They can either pass you, pass you with minor/major corrections, or fail you. I've never heard of anyone failing their defense in the last 15+ years, so don't worry too much about this. Your advisor wouldn't let you defend if they didn't think you could pass, as it is a waste of everyone's time, and it is a black mark on their record. If you have a job lined up for the fall, you particularly shouldn't feel like you're going to fail, but just be prepared. After 5-15 minutes, you'll be asked to come back to the committee, where they'll explain any corrections you should make, and congratulate you as a doctor. Pop the champagne!

Over the next 6 months, you need to enact the corrections suggested by your committee before resubmitting to the university. These corrections are sometimes as small as typos, sometimes as big as doing additional calculations or experiments. In general, you should be able to do this in a week or so of working, so I'd suggest getting this done as soon as possible after your defense date to get it out of the way. You'll also have to pay $85 to the University to submit, but you can do so electronically with a tex file, which is nice. You'll need to make sure you don't have any "holds" on your account from residual fees you have not yet paid (you can check this on SSOL). Lastly, you have to fill out some sort of NSF survey on your experiences as a graduate student. Once you do all of these things and resubmit your corrected thesis, you'll be marked as complete, and you'll be awarded your PhD imminently. Unfortunately, there are only three times of the year when you can actually be conferred a PhD: May, October and January, so your transcript won't reflect your PhD until after the next conferral date. The university will send you a letter (and email) saying you've completed all the requirements that should be sufficient for you to take a job in the fall.

Lastly, the department will do some sort of exit interview with you before you leave, to assess your experience in the department, and get feedback on ways in which to improve it for future students. This is pretty laid back, but it's worth thinking about in advance. This interview will be with one of the Directors of Graduate Studies (e.g. Kathryn, Greg or Marcel).

I personally recommend taking some time off of science before starting your new job (postdoc or not), as you'll probably be quite burned out from all of the thesis'ing. Post-docs are a productive but hard part of your career, so you want to enter them at full steam. Best to take a month or two of rest before entering into one.



Note for International students. Contact ISSO and inform them about you graduation schedule. It is important that your SEVIS paper is in order.


Creating and Maintaining Your CV