This is a guide about the many “do’s,” “don’ts,” and “oh-no-I-definitely-shouldn’t-have-done-that’s” I’ve picked up while bumbling my way through research in labs across the world. It is intended for students who are just starting out, or who suspect they might want to start out, or who have already started out and are now quietly panicking in a corner. My own adventures have mostly been in AI/ML, but the general principles apply whether you’re studying neural networks, astrophysics, or why the toaster sometimes smells like burning rubber.
Before you begin, the single most important thing to bring is an open mind and possibly snacks. Most of my best project ideas came not from grand moments of genius, but from listening to someone else and thinking, “Huh… that’s odd.” Sooooo what follows is, hopefully, a good read, though I make no guarantees about usefulness, accuracy, or whether you’ll still have your sanity by the end.
The first question I asked myself was, "Why do I want to pursue and do research?" Finding the answer to that took me almost a year, and no one's answer is the same. The main thing that made me decide I want to pursue research as a career is simply the pursuit of always finding something new. The field of computer science (AI especially) is growing at a tremendous rate. Several times I've had ideas that were already done by someone else over a year ago or even just one week ago. That pursuit of being the next and the first is what drove me to choose research. So ask yourself why you want to pursue research.
The next thing I found to be a very difficult question was, "What do I pursue research in?" For my field, there are several directions one can choose. My answer to that was pretty simple: I have always been a fan of languages, and finding out in my second year that AI models can generate and understand text was like a match made in heaven. I'd really recommend figuring this out for yourself rather than just following the latest trend. Take the time to explore, there's amazing research being done in every single field out there.
Okay, so now you've figured out what to pursue research in. The next question is, "How do I begin and reach out?" This section is very important because I'll be talking about the many struggles I faced and the things that can go wrong. Never be afraid to reach out to anyone, there is no single "cold mail template" that guarantees success. What your cold email should do is show the reader your genuine interest and a desire to learn. Remember that research is not a one-and-done thing. One collaboration could lead to a lifelong partnership. Research is also not a time-based field; I've had projects where I tried to solve a single problem for over two years, and others where I was able to solve it in six months or less. So, you will often feel like giving up, but don't. Just refresh, take a day or a week off, and come back to it with fresh eyes.
I have written over 100 cold emails to professors all around the world, and every single one was different, expressing my specific interest in their work. Did I get a response from all of them? No. One might think the success rate is higher for someone with better grades, but that's a false pretense; it's largely based on luck. I got lucky about 2% of the time, and even those responses didn't always result in an internship or research collaboration (I’ve shared examples where my emails got replies and led to something exciting.).
So now there are 3 types of mails you can write (can be sent to both the professors and their students):
The Project-Based Mail: This is the longest and often the most successful approach. Come up with your own interesting idea, write a short explanation, run some small experiments, and then email a draft to people who have worked on similar problems. This demonstrates genuine interest and initiative, and people are almost always happy to discuss a new idea with someone who has already put in the effort.
Lastly, always follow up. Professors and their students are incredibly busy, and their inboxes are often flooded with emails from students. So, when should you follow up? Give it about a week, and then send a simple reminder. Something like, "I wanted to follow up on my previous email regarding..." is more than enough and can go a long way.
Congratulations you persisted and landed an opportunity, this is just the beginning. Now comes the part where you learn as much as you can, and set yourself up for future collaborations. Here are a few do’s that I personally follow with every opportunity I get: