Three years have passed, and I have slowly realized how far back in the academic game I am compared to certain other peers around me. I knew relatively early on that I would eventually switch to industry, thus I consciously did not optimize for academic success. However, the extent of that gap is mind-blowing, and it is natural to wonder how could some people get so far ahead given a similar environment, time and resources.

But first, some meta thoughts. I did not really optimize my PhD for anything about my future, just been coasting along and trying to enjoy the research. I should obviously compare myself to those who deliberately optimized their PhD for a move into industry, rather than the academic types. I will eventually do this, as I am starting to think what to do after the PhD is over.

I did not write retrospectives for a while because I did not have interesting things to say, but recently I started reflecting on my experience and observations and (hopefully) I have something interesting to say now. You may want to read the previous retrospective I wrote (link) since I am going to elaborate on similar topics. Note that, despite being presented as a neat list of well-separated issues, all of them are interconnected.

So here are some things that held me back in my PhD, in no particular order.

Taking too long to recognize worthless ideas.

More than one year ago I noted how difficult it was for me to come up with research ideas. This is not an issue anymore and I reached the next level: most of my ideas are crap. To be honest, I do not think this is an issue in and of itself, but it starts being problematic when too much time is wasted on hopeless projects. I worked for nine months on a project that went nowhere and I feel bad every time I think at it. Because of the time wasted, not because of the project.

High performers not only have plenty of ideas, they also skim through them as quickly as possible to identify the gold nuggets. In order to do this, set up a prototype and a toy example that is as simple as possible, yet powerful enough to validate the idea. This only requires a few days to code and run a few experiments, and if they turn out to be unsuccessful learn from them, figure out what is wrong with the idea (or experiments), learn from it, and move on.

Not checking relevant literature enough.

A problem that I frequently have is that I get stuck when ideas do not work out the way I thought they should. I am slow in coming up with possible reasons and alternative approaches. Essentially, I am not very creative. And I am also biased towards (too?) simple solutions, which are unfortunately frowned upon by many reviewers. I made all of this worse by not actively seeking out what other people are doing in the field.

I am still not sure how to deal with this situation. Talking to people and reading papers can help, but is by no means guaranteed to give a solution. I never felt confident in my literature surveying skills and I am always afraid of missing that important piece of research that I look for, although it is true that the sheer amount of papers written about every topic is overwhelming and growing quickly. Yet, somehow, certain people around me seem to have a talent for finding the needle in the haystack, the one paper talking exactly about the issue at hand.

Not talking enough to people.

Being isolated was somewhat unavoidable in Covid times, but after two years virtual conferences still feel rather awkward. Very little networking happens there. I did not mind lockdowns at first because I could focus on my work without wasting time commuting, but now I realize how much I, and many others, were held back in our scientific careers by endless home office.

All researchers worth their salt have a good overview of other groups who work on the same things. This makes it much easier to stay up-to-date with the literature, as one needs to primarily be alert to what those groups publish. It also allows one to sense which direction the field is going, thus finding promising avenues for future research, which may include visiting those groups and working with them.

I avoided networking and as a result I feel like I live on an empty island. This means that there is no chance of collaboration or any kind of scientific exchange. I could have done a much better job by attending conferences and deliberately, explicitly networking with purpose, no matter how awkward and uncomfortable that is.

Not working enough with people.

This is something that I noted in the previous retrospective and I set myself the goal to establish more collaborations. I was relatively successful, as I worked on several projects with other colleagues in my groups, but I failed at the bigger collaborations with international research partners, i.e., those that are valued in an academic CV. To be fair we tried to set up collaborations with two industrial partners (not as prized in academia but much better for my own CV), but both of them failed because lawyers could not agree in a reasonable time on who owns what.

This is highly related to the previous issue, in the sense that talking to interesting people for just half an hour is almost guaranteed to generate a few promising ideas both of you are interested in.

Juggling too many disconnected topics.

Successful people tend to find a niche they like and only work within that topic and its immediate surroundings. It takes a while to get the ball rolling, getting up-to-date and publishing a few papers, but after that things almost happen without intervention. Once your name is known in the field, opportunities start finding you: people ask information about your works, giving you chances for collaborations or, at least, citations, you get invited as a reviewer, getting a sneak peek at new advancements before everybody else, or as speaker at conferences, giving you the chance of further broadening your network.

So the opposite of what I did. I published two closely related papers about vaccine design at the beginning of my PhD, then I had to completely change topic for reasons out of my control. I worked on a single project about graph neural networks, then switched to self-supervised learning, and now focusing on a mix between deep learning and statistical inference. Every time I had to start from scratch, reading hundreds of papers to get into a new field. See, it takes a while to get research-grade ideas, at first everything you think about is likely to be already published somewhere (or just senseless). So get the ball rolling and roll with it.

Not working on sexy topics with big-name supervisor.

The best way to start ahead with your PhD is to work on a hot topic under a world-renowned supervisor. Such supervisor is likely to be a well-connected idea factory and opportunities to do high-impact work would just rain on you. Simply having their name in the list of authors will increase the reach of your paper and, let’s be honest, lower the bar of publication in the top journals or conferences. This is all free attention (and citations) that you would not get, had you worked under a second class supervisor in a not-so-hyped-up topic. Citations are the currency of academia, and a good name is worth a lot more than only good research.

Not planning for the future.

There are certain, so to say, ceremonies, that are necessary for a successful academic career. For example, a research stay at a famous international lab. After knowing what you ought to do to impress this or that committee, you have to plan for it, often years in advance, and take action with purpose.

Here I am going to use Covid again as an excuse for not doing much, but for most of my PhD the future just looked like endless home office. However, I failed to plan even the simpler things like which conferences or Summer schools to go to. I did attend some, but never with the purpose and drive shown by those who know exactly where they want to go and how to get there. Truth is, I still don’t know.

Not leveraging students.

Wrapping research ideas into Master’s theses is a great way to multiply your research output with relatively little effort. Within a semester, a (good) student can go from a (good) idea to (good) results, and some additional effort on your part can result in a (good) publication. Obviously, much of this depends on the student. If you wonder how it is possible for a PhD student publish 20 papers in four years, there you have it.

At fist I was hesitating to do this, because I thought I ought to work on my ideas on my own, and maybe because I was a bit jealous of my ideas. But it was silly to feel like that, research is a highly collaborative endeavor. I started doing this only recently, but my recruiting skills are not up to par yet.


Doing good research is far from enough for a successful academic career. Other things are necessary too: work with the “right” (famous) people, work on the “right” (popular) topics, publish in the “right” venues, get as much visibility as possible, discard ideas that do not work for the first time, recruit an army of Master’s student, and so on.

All of this, and more, is what is known as the metagame. Essentially, how do you get ahead of the pack when there are no predefined rules. I have been thinking at this for quite some time, since I read this very interesting post. Here I wrote a summary of what I noticed when observing the high fliers around me and what they do to set themselves apart from everybody else in the academic game. To get good at the metagame, in general, find the good ones around you, and see what they do differently.