That's a great question. While it is challenging to be precise about what constitutes good research but I can share my perspectives. In my view, good research work starts with attempting to solve an important problem—a problem that is relevant, that people care about, one that has the potential to advance knowledge, influence practice, or inform policy.

Once the problem is defined, good research involves employing a sound methodology and evaluation framework to solve it effectively. This process includes thorough investigation of the solution space, analysis, and critical thinking.

For undergraduates, the first research problem is often provided by their professors. The expectation is for the student to demonstrate their ability to solve the problem through various tasks. Excelling in this requires several key qualities:

Taking ownership of the project: In other words, showing genuine engagement and responsibility for the project. Ability to learn new things: That is the ability to learn new concepts, tools, and techniques relevant to the research. Initiative: Proactively explore new ideas and approaches.

Having a research publication is beneficial but it is not required for good admissions. If you excel in your research, strong letters of recommendation will reflect your efforts. Additionally, having at least a technical report is very useful.

In my experience, over time, good undergraduates tend to take on more responsibility in research projects by suggesting new ideas, contributing to writing papers, presenting their work, and mentoring junior students.

What should undergrads do to gain good research experience?
To gain research experience as an undergraduate, start by seeking out opportunities within your department. This can involve joining a lab, assisting with faculty projects, or pursuing an independent study. It’s beneficial to get involved as early as possible to gain experience and develop your skills (e.g., summer before the Junior year is a good time to begin looking for projects). Finding a mentor is also crucial; a faculty advisor can provide guidance, support, and opportunities for deeper involvement in research. Be proactive in your research work by proposing ideas, volunteering for additional responsibilities, and seeking feedback to enhance your learning and contributions.

Additionally, focus on developing relevant skills such as research methods, statistical analysis, and research writing. Networking is also important; attend talks/seminars, workshops, and conferences to stay informed about current research trends and connect with other researchers. Finally, keep detailed records of your research activities, methodologies, and findings. This documentation will be very valuable when writing your statement of purpose or research summaries for graduate school applications.

Regarding building your portfolio, I’d like to request you to first pick a sport. As an exercise or training that you do everyday. Now let me tell you why. If you pick a profession or field which requires you to sit everyday, enclosed in a space, sometimes up to 17-18 hours a day, your mental and physical capacities will begin to decline. Many people start developing mental health issues, they go into depression or develop physical issues. You might graduate with a 3.97 but would it have been worth it if you also graduate with a lifelong backbone problem? This is honestly the first piece of advice I give to all my students, and something me and my colleagues also practice. Pick a game or physical activity that you like and stick to it.

Secondly, I would like to emphasize that one thing that is very hard to recover or improve is your GPA. I am not saying that a high GPA is a hard prerequisite for admission. There are countless examples of students with very low GPAs that have made into some of the most brilliant institutes worldwide. Dr. Ihsan, Dr. Zafar and myself have advised and mentored several students who fall in this category. However, It does become very difficult with a low GPA. Students often have to do some outstanding work in other components to counteract a low GPA. While of course, GPAs are relative and in some universities it is much harder to get the same GPA, you should aim for a number close to 3.5 or above. This is a decent enough number to make you competitive and you should try to remain above this number. Even if you have a lower GPA right now, you can still touch this number or come close with requisite effort in your remaining courses.

Regarding the competition in ML: As someone who teaches and conducts workshops in the field of ML and generative AI, I want you to know that the industry is still in the ‘euphoria’ phase of it all. Many companies that make LLMs for example, are realizing that a primary bottleneck is going to be network bandwidth and there is going to be a subsequent boom in Network-centered solutions too. We are still in the phase where we do not fully realize the complete repercussions of these developments. Similarly, security is going to be a big issue moving forward. ChatGPT for example, sometimes shares URLs mistakenly which leave the user prone to comm-injection attacks.Furthermore, bots are only as good as the interface they are embedded into and its level of engagement with the user, meaning HCI is also going to be essential to this boom. So rest assured, all fields are going to remain relevant in their own unique way and you should follow your passion. Conduct a small mental exercise with me. If you were at a bus stop, and you needed to go to a particular destination but all buses leading to that destination were full, you wouldn’t pick just any other random bus now would you? In a very similar manner, please do not apply to programs you think are easier to get into, but are actually not passionately invested in. You are in your prime and at the peak of your learning ability. It would be a grave mistake to believe you can later on switch easily into your relevant field after you have graduated or settled into a specific industry. I believe that if getting admitted is harder now, it is going to be even more difficult to switch fields later on, given how competitive the arena is now. Follow what you like, and the path to success will follow you too. Be good at what you do, and you will always have plenty of opportunities in terms of admissions, funding and jobs.

University selection, I believe, is also a crucial part of the process which you need to start well in advance. It is a fairly complicated process - what will be the right set of universities to apply to, given your profile. Here, you’ll need feedback from experts and experienced people who have seen similar profiles to see what will be your likelihood of getting into certain universities. Your set of universities should include places you have a good likelihood of admission and then safe universities and high-risk ones.