Failing postdoc?

Question

After a successful PhD in computer science (including some awards from top conferences, internship at top industry research labs, and a good number of citations to my ~15 papers) in a top-10 US school, I got a postdoc at a top-3 US school.

I had a number of postdoc offers and chose the one whose topic more far-related to my phd thesis, in order to learn new things, get out of the box, and show that I'm not a one-trick dog (or doc? ;) ). The topic is different but not too much far away (i.e., I didn't move from theory to systems, rather from algorithms to (applied) machine learning).

I've been in my new position almost 6 months and feel like a complete failure. I can't go on with one of the projects I was assigned to and it is not for lack of trying: I just don't understand the results I'm getting because I don't know the field well enough and I can't figure out what I am actually supposed to do to improve them.

My supervisor gets quite "adversarial" when I ask for feedback and I try to explain what I don't understand. On the other hand, he says that I'm very helpful on other projects, helping the students, and a valuable addition to the group, and he would have told me if it wasn't so. Indeed I think it is true and I'm doing a good job on other projects (which are not "mine").

The project I'm failing at could have a great impact on science (not just CS) on the long term, but I feel like any engineer trained in the field could lead it to completion, so I don't find it particularly exciting. I guess I'm missing the excitement I got when I had to prove theorems during my PhD. Right now, the project involves just messy data analysis, and a lot of try-and-error coding (mostly error), all without much feedback from the supervisor (who actually told me to ask the students...which I did and they weren't exactly helpful).

Anyway, this is getting me extremely stressed (I actually started seeing a therapist about this), and I'm thinking of moving away from this position, although I may have the possibility of renewing for another year (if my supervisor would even still consider the option, which he gave me when I started)

What are the pro/cons of moving away after 1 year with not exactly much positive work done yet, especially of which I could claim ownership? I believe that right now my supervisor would not, in the future, write me a very positive recommendation letter, in my opinion.

Note that I already have offers for next year, on topics more related to what I did during my phd.

Answer 1

Not really an answer, but some advice: it's normal to need time to get going as a post-doc. Usually you end a PhD at the end of a fairly large project. You have completed what you've been building up to. Therefore you need to start pretty much at the beginning with new projects as a post-doc, and it takes time before they come to completing. After that you'll usually have different projects at different stages, so the start of post-doc time is likely to be unusually slow results-wise.

In terms of the question: the main problem with trying to move when you've not produced much recently is in not being able to secure a job. If you've already got offers, you might want to think more about the personal side of moving jobs.

PS. You might like to look up 'imposter syndrome'.

Answer 2

You specifically asked about the pros/cons of leaving after one year without much to show for, so I will start with that. In my experience, you build credibility over time. And the more of that your resume shows the better.

Imagine the following situation:
You leave this position now and move to another school. Likely not in the top-3 anymore. Your next university takes you, after all you come from a top-3 school, even without published results one could assume you profited from that. But now, after another year, funding completely dries up at your new university and you have to leave again. Now you already have two very short post-doc employments on your resume. It might start looking like a downward spiral. If you decide to leave academia then, you would have to explain why it took you two years at different schools after completing a PhD to figure out that you were not cut for academia.
I am not suggesting that this is bound to happen. I am just pointing out a negative scenario to consider. And there will always be people who made something like this work.

An argument in favor of leaving would be, if you cannot at all imagine turning the situation around, "wasting" more valuable time and ultimately having nothing to show for after 2-4 years - or however long you stay.

---

But aside from the pros/cons, I believe a careful examination of your situation is called for.
You say, "I just don't understand the results I'm getting because I don't know the field well enough and I can't figure out what I am actually supposed to do to improve them.", however, later on you point out: "any engineer trained in the field could lead it to completion, so I don't find it particularly exciting. I guess I'm missing the excitement I got when I had to prove theorems during my PhD."

With regard to the struggle of not knowing what to do, I would respond: You have a PhD! You should have acquired the ability to do a thorough literature search and familiarize yourself with almost any topic within a reasonable amount of time. Especially, if your new field is not too far from your old one. Read publications, text books (if they exist) and talk to others in your new field. Is your new supervisor the only faculty member at your top-3 school that deals with this matter? Or are there others with a similar interest? After six months, you should have gained some understanding of what you are doing and what the results mean.

This brings us to the second part, the lack of excitement. As a general advice for career choices, I would recommend putting down a list of expectations that you have with regard to what you do. And then assign each item a priority. But only assign each priority once. Based on your question, this list could contain:

• I want to conduct research at a top-3 university
• I want to conduct research in a field with which I am familiar
• I want to conduct research in a field that I find exciting

Once you have assigned a priority to each of those, only one has the top spot. Follow that top priority. If you current position does not satisfy that, move on.

Answer 3

If you've realized after a year that you want to go back to doing theory research, you should simply do that. It's not "bad" to leave a post-doc after a year and take a different one (I did that, though for different reasons). When you apply for permanent jobs, if you don't have a letter from your current host, that should be fine as long as you have good letters overall, including one from your PhD advisor. Your overall record will matter a lot more than what project you did where.

In trying to figure out what you should do, I would suggest simply figuring out what you want to do. It is sometimes possible to "overstrategize."