There seems to be a lot of information out there about preparing for the preliminary exams, and for the qualifying exams, but virtually nothing about the job application process at the end of one's Ph.D.—perhaps this happens because those who have passed the aforementioned exams are still in grad school, whereas those who have obtained a job have moved on elsewhere, and just didn't think about contributing information about the process to the MGSA. Also, it seems that in general, mathematics departments do not have much information about applying for jobs; this is in sharp contrast to, for example, the economics departments at top schools, which are well-oiled machines in terms of preparing their students for the academic job application process.
The hope is that this entry will eventually rectify this situation for those graduating students who hope to obtain a postdoc at a research institution after their Ph.D., and recent and current Berkeley postdocs are strongly encouraged to contribute. Also, the hope is that soon, there will also be pages up discussing other application processes—e.g., for tenure-track positions at smaller schools, and for non-academic jobs.
As the definition suggests, a postdoctoral position is one taken after you finish a Ph.D., and before you take a tenure-track assistant/associate professor position. Postdocs are usually at universities, but sometimes at other institutions with research groups (e.g., Lawrence Berkeley, Microsoft, IBM). They are temporary, non-tenured positions, and in mathematics, most postdoc positions last 3 years, but sometimes as short as 1. In most non-scientific fields, graduating students typically apply for tenure-track positions in their final year, but in the sciences, postdoctoral positions have nearly become a prerequisite if your eventual goal is to obtain a tenure-track position at a research university.
If your goal is to obtain a tenure-track position at a smaller school (e.g., a liberal arts school), then doing a postdoc isn't necessary, and you can apply for these jobs straight out of graduate school. The job application process for these two types of positions is very different. For tenure-track positions at smaller schools, the common route seems to involve attending the January joint AMS-MAA meetings, plan to have in-person interviews there with various schools, and if that goes well, to be invited to give a "job talk" (i.e., a presentation on your research) at schools. For postdocs, the application is more or less like graduate school (although, as discussed below, there are certain other "soft" things you can do to improve your application): you write up an application about yourself and your work, and send this, along with letters of recommendation and your academic record, to a school.
The best place for job listings in the US is the Employment section of the AMS website, and the affiliated MathJobs application database. The Mathematics Jobs Wiki has a useful list of postdocs of interest to research mathematicians, and this wiki also has a list (currently out of date).
Also, try the departments' websites. These will typically have a page for job listings. If you're very interested in a particular school, it's good to check their website, just in case you somehow missed their listing on other sites (or for some reason they didn't list).
The NSF fellowship provides two years of full-time support for recent PhD's. Most are awarded to students when they graduate, but postdocs are also eligible. It is only open to US citizens. About 30 fellowships are awarded every year, and in 2007, 4 students from Berkeley received the fellowship, which is a fairly typical number. The stipend is around $60,000 a year.
It is possible to take fellowship in chunks as short as a month at a time. It is also possible to take a half-stipend for part of your fellowship and use a semester of fellowship over the course of a whole year. It is common to do this while obtaining part-time support from another source, often the institution where the fellow is.
Every fellow has to have a host institution and "sponsoring scientist" (a mathematician at that institution). These are specified at the time of application (the fall before beginning your fellowship), and will be taken into consideration when evaluating your application. It seems to be conventional wisdom that choosing a slightly less prestigious host institution will improve an application's chance, though as can be seen on this list, several awards were given out for MIT, Harvard, Stanford and Berkeley in 2006.
You should inquire with potential sponsoring scientists well before the deadline, since they need to write a statement from your application (it's not a letter of recommendation; you will be able to read it and suggest changes, and you upload it to the NSF, not them), and may need to consult with deans. This would also be a good time to ask about things such as whether you can teach at that institution. Just like when asking for recommendations, it's a good idea to attach a copy of your CV to the email you send them.
The guidelines for the NSF Postdoctoral Fellowship can be found here.
Information about NSERC Postdoctoral Fellowships, which are for citizens and permanent residents of Canada, are available at . Information about Fulbright postdoctoral opportunities can be found here.
Many universities, especially the more prestigious private schools, hire a number of postdocs every year out of their own funds (unlike, say, Berkeley, where most postdocs are supported by a grant or external funding). For reasons that remain unclear, these positions are typically named after a important mathematician in the pre-history of the relevant department. These jobs are usually 2 or 3 years, and typically involve a relatively light but non-zero teaching load. Increasingly, applications for such postdocs are being handled through Mathjobs, though a few notable exceptions, such as Yale require you to apply through their website.
The NSF is currently funding (partially or entirely) a number of postdoctoral positions through the RTG programs (VIGRE has been discontinued). These are basically indistinguishable from the "named" postdocs mentioned above (and will probably be advertised through MathJobs), except that they are only open to U.S. citizens and nationals, have slightly lower teaching loads (NSF guidelines say one course per semester), and will be focused on a specific sub-area of math.
There are eight NSF-funded institutes which offer various post-doctoral positions, though several of them for a shorter period (one semester/one year), and sometimes only in conjunction with thematic programs:
Most graduate students look exclusively at positions in the United States. However, there are many great opportunities available in Europe and Japan which are open to US citizens and don't necessarily require knowledge of the local language. Note that the job cycle in other countries may run on a different timeline than in the US.
In the British Isles, the London Mathematical Society maintains a list of job offerings. Increasingly, these jobs are also posted on MathJobs.
Institutes analogous to the ones above overseas include:
One would like to believe that what matters most in the end for one's postdoc application is the quality of the thesis work, and how significantly it contributes to mathematical knowledge. While your research is the most important thing, there are certain little things that a graduating student can do to make themselves and their research better known. You are much more likely to get an offer from a university where someone knows you and your work, than if no one knows you at all. As such, here is some advice about how to get your name out there:
William Stein, a former Berkeley graduate student, has posted his application materials (both from his first job search after graduating in 1999, and from a tenure track search in 2004) here. Ben Webster, another Berkeley PhD, collected a bunch of application portfolios.
For academic jobs in math, the cover letter is something of an antiquated formality, with most of its functions taken over by the AMS cover sheet. However, it is explicitly required for many jobs, and it's a good idea to include one even if it isn't included in the required materials.
The cover letter should:
It's a good idea to use UC letterhead when writing cover letters. The easiest way to do this is using the
ucbletter class in LaTeX. You can see an example cover letter in this format here.
The TeX FAQ has an entry listing useful packages for creating a cv. I found currvita useful.
Having good letters of recommendation is one of the most important parts of one's application. Especially if you haven't published yet, the hiring committee will be relying on the recommenders to get an idea of your potential. As such, it helps a lot if (1) the recommenders are familiar with your research, and can comment on it; (2) the recommenders are important people in your field.
Remember to give your recommenders several weeks notice before your first deadline, as writing a recommendation letter is a somewhat time-consuming task and your recommenders are probably busy people. Still, don't be afraid to apply gentle reminders until you're sure they've gotten the letters everywhere they need to be; usually online submissions to MathJobs and the NSF are sufficient, but you can arrange for them to give hard copy to Barb Waller which she'll copy and send out if any hard copies are needed.
It seems to be the general consensus that you should have a letter of recommendation from a non-Berkeley person. You don't have to know this person very well; the most important thing is that they know your work well. Obviously, someone you have a closer relationship to is preferable.
Remember that for most postdocs, you will need to have a letter of reference that deals with your teaching. It would be best if you could get this letter from someone who has actually seen you teach. So, it's generally wise to ask any professor you teach for to come and observe you for a class. If that professor isn't willing to, ask the graduate vice-chair (currently Daniel Tataru), your advisor, or another professor.
Although most postdoc applications ask for three letters of reference on your work, it is not a bad idea to have four or more, if there are enough people who can write good letters for you. The reason for this is simply that more good recommendation letters means more positive things said about you to the hiring committee.
When multiple students of the same advisor, working in the same area, graduate at the same time, there is the risk that they will compete with each other. In such a situation, it is not uncommon for the students to coordinate with each other and their advisor, and try to avoid too much overlap.