Kira Goldner
About
Research
Teaching
Diversity & Service
Resources
Blog


The Strategy Space

Blog home »

On Postdocs

June 13, 2024

Many people say that a postdoc should be the best time of your academic career, and they're half-right. But there are a lot of things about being a postdoc that are extremely difficult that aren't widely talked about. I think a lot of these can be mitigated with awareness and planning.

Index


The Transition

Transitioning from your PhD to a postdoc is hard. You're going from THE most confident you've ever been academically, having hit your stride, in a place with all of your connections and support, having found your topic, to somewhere completely new.

And, a lot of people see the postdoc as an opportunity to really try out new things, adding even more to the newness. New research directions, new collaboration relationships, etc.

At the beginning of your postdoc, you basically feel like a first-year PhD student all over. Yet, you expect yourself to be performing the way you were as an nth year PhD student with all of the momentum you had in your peak, after you'd established yourself and built all of your previous connections. And there's all of this pressure to produce, to do things that are different than your PhD, and quickly, because you'll go on the faculty market soon.

So what can you do about this?
  1. Expect to feel this way.
  2. Cut yourself some slack. It's basically like starting over as a first year, except that you already have experience, so it will go much faster this time. But, you will NOT feel like you currently feel, with all of your comfort and connections.
  3. Don't go in too many new directions, especially at first. Pick something close to your area of expertise so that you can get your comfort and confidence back as you build up your new home and connections.
  4. Find a way to hit the ground running. If you can make one of your projects with people you already work with, or start working with your new people in advance so that you feel more comfortable and can hit the ground running, that's a good thing. It's the newness that can be a bit limiting.
  5. Give yourself whatever comfort you can. If this means a longer commute so that you live closer to friends, think about that, because a lot of things will feel new and uncomfortable, and you don't have a long time to get settled. Or, if having a short commute makes you more comfortable, do that. You don't want to change all the factors in your life at once. Change kinda sucks. Minimize it.
  6. There's a lot of pressure to produce quickly during your postdoc, as you have to find a job soon, but the good news is that there's usually carry over from your PhD, so you don't have to worry quite as much about that, as that will keep you looking productive for the first while.

The Purpose of the Postdoc

The most common aims of the postdoc are: To this last point, I've specifically heard, "One of the aims of a postdoc should be to get a rec letter from someone famous at a different school than your PhD." It's definitely a good thing to think about, but if you already have strong letters from a variety of people (including some well-known), including some outside your PhD institution, then this doesn't need to be your top priority.

Similarly, if you already have breadth in what you work on, or already have clear independence from your advisor, or already have a great body of work, these don't need to be the aims of your postdoc.

Postdoc Types, Funding, and Applications

There are three types of postdocs: Industry postdocs, centralized postdocs (e.g., with fellowships), and ad hoc postdocs informally arranged.

Industry and centralized postdocs both require formal applications with deadlines around December 1. The only industry positions I'm aware of are at MSR, Amazon, and IBM.

The following are some centralized positions that are most relevant for those in theoretical computer science.
These centralized postdocs are more "prestigious," sometimes give you the flexibility to work with multiple people instead of the single person funding you, sometimes come with a travel or equipment budget, etc. They tend to be resolved around January and February. Applications will require a research statement, CV, ~3 letters of recommendation.

Ad hoc postdocs are arranged informally by email. You or your advisor should email the person you're interested in working with. These arrangements tend to form later, after the centralized postdoc market has cleared.

At the beginning of your postdoc market year, you should look to start advertising yourself. Some communities specifically make lists of people going on the market, including the postdoc market. For example. SIGecom Exchanges organizes "job market profiles" every year, and I know I always check them to see who's on the market. Some conferences organize "Graduating Bits" where those near the market give ~3 minute spiels to advertise themselves, such as ITCS and EC. Make sure you participate in these opportunities.

One thing to keep in mind: as a faculty member, it's very difficult to get one's hands on postdoc funding. If someone tells you that they don't have funding for a postdoc, do not be offended. It is far rarer to come by than student funding (at least in the US), and even if someone does have it, they might only have enough for one postdoc in the next five years and are saving it for someone. It doesn't mean they don't like you and wouldn't love to work with you. It much more likely means that postdoc funding is scarce.

To this end, if you can find your own funding (by fellowship), you can substantially open up your opportunities here. I don't know of that many options, but one, for example, is the NSF Math Sciences postdoc fellowship.

Another option is two people sharing a postdoc (funding-wise and research-wise). This might be two people at the same institution, or could be two people at nearby institutions. For example, my colleagues at BU often share postdocs with colleagues at MIT, Harvard, or Northeastern. On paper the postdoc usually spends one year at each university to simplify funding, but often works with both faculty the whole time. I've even seen longer-distance postdoc shares for a year here, a year there.

Postdoc salaries are highly variable. Most recently, I've seen them vary from $65k to $105k, where the high end requires teaching a class each year, and about $15k/year of that is typically paid by the institution for teaching (making it less onerous on the person with limited postdoc funding to actually have that much for you). Sometimes there are equipment and/or travel budgets, usually maxing out around 5k for the postdoc. Industry postdocs, of course, pay much higher.

If you are looking for details on application materials, you can see my existing post on the faculty job market which includes materials there. I also plan to write a post at some point in the future about applying for fellowships, which will include my successful applications for some of the above mentioned postdoc fellowships.

The Length of the Postdoc

Computer science postdocs are typically 1-3 years. Which length should you go for?

1 year postdocs are only feasible if you go on the faculty market and postdoc market at the same time and then defer your faculty position for a year. You absolutely cannot get a postdoc and then plan to go on the faculty market during that year. You will spend the entire year dealing with the faculty market, so your postdoc will essentially not exist. It will be a waste for you and for your postdoc host. If you do already have a position, 1 year postdocs are great. They're low key, no pressure, since you already have a position, and a great time to grow. Note that I have never heard of a US university agreeing to deferring a faculty position by more than 1 year, and only rarely (so far once) for non-US faculty positions.

In computer science, 2 year postdocs are the most standard. They're also the reason for the pressure described above. You essentially have one year to transition and produce, and then a second year to primarily focus on the faculty market.

A few positions are 3 years long, although these often require teaching. 3 years are nice because it's less pressure to quickly produce and get on the market. However, teaching is a huge time-suck, and one of the arguably best parts of a postdoc is that you can focus solely on research. This may be appealing to some people who are very passionate about teaching, or who already have materials prepared, or who want to prepare them in advance of their faculty job.

Which Search? Postdoc or Faculty?

If you're considering a full faculty search straight out of your PhD, one of the primary questions is whether you're a strong enough candidate without a postdoc to get the position that you want, and that I cannot answer for you, but some of these other points may be relevant.

One question I hear a lot from people coming out of their PhD is "should I just do a limited faculty search now?" Sometimes people don't want to do the full faculty market, or they hear about an interesting position.

You should only do a limited faculty search if the schools you're applying to include your absolute top choices. If you do a small search, you're cutting down your options. If you hear about specific positions for exactly what you do at your top one and two choice schools, then sure, apply. (When schools DO specify, it's usually because someone left them recently and they're trying to fill the void, and so they might be rushing to do that. But different schools are different.) But if schools are just generally hiring, then they will always be generally hiring. Going on the job market is a really big thing. Even if you only apply to a few places, the preparation of materials, job talk, and the whole experience is totally exhausting and will take all your focus. Of course, there's so much randomness every year that it increases odds to throw your hat in the ring potentially multiple times, if you are willing to trade off that focus/energy/everything. But if you're going to throw your hat in the ring, don't limit your search unless you're absolutely happy with those limited options.

Another thing to consider is that in some ways, being faculty kind of sucks. The first 2+ years of the job, a lot of what you is do getting things in order, administrative nonsense, and teaching, just looking for a moment to think about research. During a postdoc, you literally have nothing on your plate except to do research (although obviously the pressure of eventual job-seeking after). There are, of course, benefits to being faculty, but you'll never really get to focus in on a problem again for a long time. You have about one hour a week to think about it, and then you advise others on how to do the leg work, but you don't get to. So consider which you want to do now when you decide what to search for next.


Thanks to Shivam Nadimpalli for many of the conversations off of which this post is based and edits on the post!

Blog home »