How Does a
Regular Student
Get Into Research?

If you're a first or second-year student in a great, or not so great, engineering institution in India, this blog is probably what you wanted to know about if you're interested in the scary and exciting world of research. I am going to attempt to demystify this process, give you pointers on how you can hopefully get into it, and maybe even attempt to climb the upper echelons of academia within and outside of India. And before you ask — no, you don't need to be from an IIT, and no, your CGPA probably isn't as important as you think it is.

01 // Let's kill the myths first

There are four beliefs that circulate in Indian engineering colleges about research, and all four of them are wrong. I want to address them directly before anything else because if you believe any of them, they will stop you before you even begin.

The first is that you need a CGPA above 9 to be taken seriously. The second is that research opportunities only exist for students at IITs, NITs and a handful of other institutions that everyone has already heard of. The third is that you need to already understand a research area deeply before you can approach anyone about working in it. The fourth, and perhaps the most quietly destructive, is that research is only relevant if you're planning to go into academia.

All four are false. The CGPA belief stops students from approaching professors who would happily work with them. The institution belief causes students at perfectly capable universities to never explore what's available to them locally. The knowledge belief creates a catch-22, because you can't learn the field without being in the field. And the academia belief causes engineers who would make excellent researchers to redirect themselves entirely into software jobs without ever testing whether they might have loved the other path.

Research is a way of engaging with problems that don't have clean answers yet. That skill is useful everywhere, not just in universities.

The library is where it starts — and where it gets real.

02 // How my research actually started

My own entry into research was about as unglamorous as it gets. A professor mentioned at the end of a class, almost casually, that she was looking for students interested in working on projects. Not a formal announcement, not a competitive application process. Just a question thrown into a room.

I had no idea what research actually meant at the time. I did not know what a conference was. I did not know what a citation was, or what it meant for a paper to be published in a journal, or what any of that implied for a career. I just said yes because it sounded interesting. My co-authors (fellow classmates) and I were in the same position. We knew nothing.

That is genuinely how most research starts. Someone asks if anyone is interested, and one person says yes.

03 // On the CGPA question

CGPA matters significantly less than the culture around it would have you believe, at least in the context of undergraduate research at a typical Indian engineering institution.

Here is something worth understanding about how universities actually function. Professors are not primarily teachers. Teaching might be the least important thing in a professor's job description, which is also why some of the most intellectually formidable professors you will encounter are genuinely poor in the classroom. Their training, their incentives, their passions and their professional reputations are all built around research, not pedagogy. They are evaluated on the papers they publish, the grants they secure, the students they mentor into producing significant work and the conferences they present at. The classroom is, for many of them, an obligation they fulfill alongside their actual work.

What this means in practice is that most professors at most Indian universities, outside of the very top institutions where competition is intense, are not turning away student collaborators. Many of them are actively looking for students who are curious and willing to put in time. The bar is often not your CGPA. It is whether you seem genuinely interested and whether you can be useful.

The number matters more in specific contexts. If you are applying to work with a very well known professor who has many applicants, or if you are applying for a formal research program with a competitive selection process, CGPA becomes a screening tool simply because there are too many applications to evaluate individually. But for the vast majority of undergraduate research opportunities, especially the ones that students from average institutions can realistically access, you should not let a number be the reason you never knock on the door.

Ask and you shall receive far more often than you expect.

04 // Cold emailing professors

Cold emailing works. I say this not just from general advice but because I know people personally who secured research opportunities in entirely different countries while doing their undergraduate degree in India, purely through cold emails sent to professors they had never met. It is a real strategy and it is underused because most students assume it won't work and never try.

But it depends almost entirely on the quality of the email, and most cold emails are not good.

The generic email does not work. If your message could have been written by any student and sent to any professor, it will be treated accordingly. A professor receiving a cold email can tell within two sentences whether the sender has actually read their work or is simply mass-mailing anyone in a vaguely relevant department. The former gets a response. The latter gets silence.

At NYU, and at many research universities, the more common pathway is different. You take a professor's class, you perform well, you engage in office hours and ask good questions, and opportunities emerge organically over the course of a semester. That pathway exists because the professor has had time to observe you directly. The cold email is essentially trying to accomplish the same thing in writing, which is why the bar for the writing is high.

If you are an undergrad in India with four years ahead of you and no shortage of time, the cold email is absolutely worth attempting. Start with professors at your own institution because the bar is lower and the access is easier. Then expand outward. Professors who are not well funded, which is the majority of professors at the majority of institutions, are always looking for motivated volunteers. The worst that can happen is they say no or don't respond. That costs you nothing.

One specific, well-researched email beats a hundred generic ones.

05 // Learn to read a paper. Seriously.

This is one of the most underrated skills in all of engineering and it will pay off regardless of whether you stay in research or go into industry. Knowing how to extract value from a research paper quickly, how to identify what the actual contribution is, how to evaluate whether the methodology holds up and how to trace an idea back through the citations that informed it is genuinely useful in ways that go far beyond academia.

Even in industry, particularly if you're lucky enough to work on problems that require actual thinking rather than just following instructions, you will sometimes encounter a situation where the answer to your problem exists in a paper. An obscure paper, possibly in a language you had to run through a translator. Knowing how to navigate that literature, how to find it and how to read it, is a meaningful advantage.

There are a few tools worth knowing about early. Google Scholar is the obvious starting point for finding papers. arXiv is where most researchers in machine learning, physics and adjacent fields post their work before it goes through formal peer review, sometimes years before it appears in a journal. This means you can read what the actual frontier looks like right now, not what the frontier looked like when a textbook was written. Semantic Scholar and Connected Papers are useful for understanding the relationship between papers and finding related work you might have missed.

Understanding the structure of academic publishing is also worth the time. In computer science, conferences are often more prestigious than journals, which is the opposite of how most fields work. Getting accepted to NeurIPS, ICML, CVPR or CHI is a significant achievement that carries more weight in the ML and HCI communities than most journal publications. Within conferences there are also workshops, which are smaller and more informal and often a more realistic entry point for students presenting early-stage work for the first time.

Research is a team sport — even when it doesn't feel like it.

06 // The part nobody tells you about

The hardest part of research is not finding a professor. It is not writing the paper, and it is not getting it accepted. It is motivation, and specifically the kind of sustained motivation required to keep going when the work is slow and unglamorous and not producing results.

Starting a research project often feels exciting. There is a problem, there are ideas, and there is the appealing possibility that you are going to figure something out. What follows is often many hours reading through existing literature before you can write a single line of code or run a single experiment. There is waiting for access to compute resources, or the right data, or the right conditions. There are experiments that do not work the way you expected. There are results that are interesting but not interesting enough. There are reviewers who will read months of your work and respond with comments that take another few months to address.

This is not a discouraging reality. It is simply the reality. Research moves at a different pace than coursework, where every assignment has a deadline and every correct answer can be verified immediately. In research the feedback loop is slow, the validation is uncertain and the reward is often delayed by a long time from the effort that produced it.

That quality, the ability to keep going through the unglamorous middle part, matters more than almost anything else and almost no one talks about it.

Go send the email. The worst that happens is you learn something.