The Rabbit Hole

The next part in the series is a dedicated feature from a dear colleague of mine from undergrad! They’re a very successful scientist now, someone who I deeply respect, and a colleague who I saw in action first-hand during undergrad. Now why the preface? Because even though I only expected rather brief responses to my call for tips, they came back with a detailed and well thought-out set of responses. In fact, they wrote enough for a whole post and inspired me to continue this as a series! Their writing alias will be “Deckard”, as in “Deckard Cain” from the Diablo series because they embody the famous quote: “stay a while and listen” (and not the whole killed by an allergy to a butterfly thing). I’ll sometimes add a comment at the end so that I can act like we’re in an undergrad lab again.

Tip DC 1: Write a grant

This isn’t something that only professors do. Whether you stay in academia, found a startup, or find yourselves in industry working in a large company or research lab, writing research and funding proposals will be a significant part of your post-PhD life. One of the expectations of someone with a PhD is that they can organize and lead research projects. This doesn’t have to mean projects that end in journal papers.

For example, startups apply for grants to demonstrate minimal viable products, and matured companies with tens of thousands of employees submit proposals to win contracts that are worth multiple millions or even billions of dollars. Most companies and government labs also maintain internal funding (and application processes) to advance promising ideas. Therefore, it is imperative that you gain experience in writing proposals. There is no better place to do this than academia. Apply to scholarships and travel awards. Ask your PhD advisor if you can help them write an upcoming grant. They are highly unlikely to decline. Beyond the experience of planning a multi-year research project, you’ll get good at selling an idea. If you get good at that, you can literally create your own job prospects out of thin air. Two separate times in my career I applied to companies that weren’t even hiring. But I approached them with an idea and a grant that we could apply to for funding. I landed one internship and one job that way.

Ben’s comment: This is great advice! Being able to sell an idea and to express it coherently is an important skill to have. Some people may excel doing so orally, but may be weaker writers – it is important to develop both of these skills for academic funding and future career opportunities!

Tip DC 2: Stay in shape

There will be all-nighters. There will be long days. Some days you might have back-to-back meetings and lectures (with you doing the lecturing) and a deadline coming up. You may have to sit through a 6-hour qualifying exam. Getting through these and still having the energy to work on a speculative idea are feats of physical stamina as much as mental stamina. Eat well, exercise regularly, and practice good sleep habits so you have the energy to power through those busy times and not get totally burnt out.

Ben’s comment: On that note, invest in an ergonomic setup for your desk(s). Also eat well. Junk food will satisfy your 20’s palate and stomach in the short term but your imminent transition to your 30’s will hate you.

Tip DC 3: Take a probability and statistics class

This one comes from my experience in the engineering world. In school, the problems are structured like: here are X, Y, Z, find A. But real life is messy. Equations or models are essentially always approximations. The inputs to them are noisy estimates. In real life, the problems are more like: here are possible values for X, Y, Z, find the possible values of A. Any “point” estimate for A is practically useless unless you can also quantify how far the real value will be from your calculation.

Let’s put it another way: good design is about managing uncertainty. How you describe uncertainty in a rigorous way is with probability. It is easy to avoid stats classes both in undergrad and grad school, but taking even one graduate-level stats class will be of tremendous benefit to you. You may be able to avoid probability and stats in school, but not afterwards.

Ben’s comment: For any of the STEM nerds out here, take this advice to heart. I am personally learning (the pain of) this right now. Stats is a very different way of thinking than first-principles approaches, but are often the only (useful) tool we have in industry (e.g. data science, AI/ML, finance, etc.).

Tip DC 4: PhDs are not over-specialized

“I don’t want to get a PhD because I will end up with a niche skillset and pigeonholed into a particular type of job” – I have only ever heard this opinion from undergrads. Compare the first 5 years after undergrad for someone hitting the job market versus someone going to grad school. With only a Bachelor’s you will go to work at a company. You will learn to use that company’s software working on some problem. You will get very good at it, making yourself and the company money. It is not in their interest to move you to a different project or area and slash your productivity while you learn a new skillset. After 5 years you are very good at a particular problem with a particular software.

Grad school is different. You take all kinds of classes. You work on a research problem and attack it with various methods only to find that most of them prove fruitless. Maybe all of them prove fruitless and you switch to a different project. Eventually something works and is novel enough to win you a PhD. After 5 years you are familiar with a catalogue of analytical tools.

Here are some examples. You’ve been going to research seminars and academic mixers for years. Most of the techniques you’ve only dabbled in or heard about at talks, but you know they exist and what they might be applicable to. You know what to read in order to learn more about them. Any problem can be thrown at you and some aspect of it might jog your memory about “that time you tried to fit the data with a conjugate prior” (and you know what that is because you took a statistics class!). It didn’t work out then…but now it just might. Unlike a company boss, it was very much in your PhD advisor’s interest to expose you to a broad spectrum of research projects and methods in the hopes you might stumble upon the combination of the right tool and the right problem. It didn’t matter that you spent the first 2-3 years finding nothing but dead ends, because no grad student gets credit for doing something that’s already been done.

Now let’s turn to industry; think about what happens when you get laid off. If you’ve been working at a company, the other companies might not use the same software that you’ve just spent 5 years learning. They might not even be doing the same thing. You never had a chance to branch out and explore dozens of ideas that didn’t work because failing costed money, especially when you were getting paid the big bucks in industry and not a grad student stipend.

The moral of the story is this: don’t listen to undergrads about over-specialization. Get a PhD.

Ben’s comment: to add to this, over-specializing in something tends to imply (to others) that someone is good at “nothing else but”. This can be true, but poorly represents those that have a PhD. We are typically poor at marketing ourselves and our skills because we have never really needed to do so (see tip DC 1) – our grades and research outcomes have typically done the talking. So to this point, those who are looking in from the outside may think of grad school as refining a singular, niche skillset. However, it is not! Grad school (and getting a PhD) typically requires that a very well-rounded skillset be developed so we can support research into a singular, niche topic.

Make sure you become aware of your skills and work on advertising yourself! Realize that even though the “over-specialization” sentiment can often be held by undergrads and those outside of the academic environment, the vast majority of the world consists of those people (including recruiters).

Tip DC 5: Keep a notepad to jot down notes by your bed

A simple one now: go to bed with a notepad by your side. The best ideas bubble up from your subconscious when you are drifting off in thought about something else. Finding good ideas can be like looking for something in the dark: often you can only see it in your peripheral vision.

Ben’s comment: I find this to be true, but I also found it to be somewhat disturbing when trying to be well-rested. It’s probably an issue of my own as I am a nervous lil munchkin.

Tip DC 6: Unproductive time can be productive

You’re in your office or at home. You’ve been staring at a blank page or code editor for 30 minutes and the words or code just aren’t flowing. So you decide maybe this just isn’t the day and go do something else.

That would be a mistake.

When working on a research problem you can’t measure productivity with lines of code or words of text per hour. Some days you’ll get 5 minutes of work done in a whole day. On others you might achieve 6 months of progress in a weekend. It might not feel like it, but any time you are focused on a problem you are generating fragments of ideas for your brain to chew on. If you break focus just because you haven’t written anything down, you interrupt the process. At the end of the day, you still might not have written anything. But one evening a week from then you might have a sudden insight you may not have had if you’d judged your progress by your word count.

Ben’s comment: certainly true! I naively thought that being an experimentalist would release me from constantly thinking about my problems away from the lab. However, I found that whenever I was stuck, I was thinking about the problems as I commuted or in the shower. While I physically removed myself from the problem, I found it fruitful to keep the gears going when doing something monotonous (e.g. a walk, shower, etc.).

Tip DC 7: When writing, stop writing when you still know what to write next

You spend a lot of time writing in grad school. You write code, you write papers, you write reports, and eventually you’ll write a small book for your PhD dissertation. At the end of each writing session make sure to stop when you still know what you will write for the next paragraph. It will be much easier to begin writing again at the start of the next session compared to if you had stopped at a more natural break point.

Also, I find it takes about 30 minutes of staring at a blank page before the words start to flow. Expect this, plan for it, and definitely don’t get up because you haven’t written a thing in 29 minutes.

Ben’s comment: this resonates with me a lot because I recently finished my thesis. If you are finding yourselves having a hard time starting, try this out! Starting and getting into the flow is often the hardest thing – once you start, you stop doubting yourself and are able to express your ideas much more fluidly. This is why it’s nice to brainstorm with others “to get the juices going” in projects, and also why it’s nice to have what I call “anchoring ideas” when explaining things (e.g. at your defence).

So that’s that! A sincere thanks to Deckard for their contribution because they offered a lot of insight (and foresight) into grad school and post-grad life. One overarching theme that I noticed while reading Deckard’s response is the importance of personal development alongside professional/academic development, although its separation can be fuzzy. Grad school is a significant character-building journey that tests not only your technical abilities, but also soft-skills such as stress, time, and project management. And this is a bit meta, but you will also practice mentorship skills as you become a senior in your research group. As Deckard demonstrates nicely, mentorship requires being self-aware and able to communicate what you learned to others.