Sushant Kumar - Harvard University

Sushant Kumar (EE1) Domain: Research-Machine Learning

Interned at Harvard, Massachusetts

Mode: Online

Interviewed by: Samarth Tiwari

MOTIVATION

Even before cracking IIT, I used to read articles about new technologies being developed and research going on in top colleges of the world. I have a keen interest in Applied Mathematics, Quantum Computing, Machine Learning (ML), and Large Language Models (LLMs). After the boom of chatGPT, I became fascinated by the applications of Machine Learning. So, I wanted to pursue a research internship in ML. I got selected to be a part of the HW1 group (Hardware Dataset 1) at the Edge Computing Lab at Harvard. I was involved in the implementation of the frontend and backend of RTL Design. I also worked on the integration of silicon compiler and Langchain prompts. 

PREPARATION AND APPLICATION PROCESS

I like to read research papers and articles. Google DeepMind was a handy resource. It has lots of articles and podcasts about recent technological breakthroughs. I used to spend time reading primarily on weekends, so managing coursework was also easy. I would clarify that a sophomore is not expected to be an expert in any field. You just need to show your willingness to learn and be excited about the project. If someone wants to dig deep into any topic, I  would suggest approaching a professor on campus as the best way to get started. Coming to my application process, cold mailing was the way to go. I wrote around 200 e-mails to different professors around the world. I was seeking international opportunities, as I feel that I will have ample opportunities to do research in India as well. I want to be in a position to compare both and make future decisions based on that. I started emailing professors in January. My CV was shortlisted at Harvard, and I was sent research papers to read. I was then given a dataset based on the research papers to check my understanding and application skills. My math skills proved to be advantageous and helped me perform well in their evaluation and around February end, I got the acceptance letter. Following my acceptance, the professor sent me 1-2 research papers every week. 

INTERNSHIP: WORK AND EXPERIENCE

The research work was from May to July end in online mode. For my part, I used to get a research paper every week and had to work on its implementation. The daily work was for 1 to 1.5 hours, but the working hours were flexible. We had weekly review meetings on Fridays. Surprisingly, the approach to the work was something very unexpected for me. I thought being well-read about a topic would suffice, but in reality, you need to stay updated about the latest developments in the field. My instructors during the research were very knowledgeable and helpful. They provided me access to Harvard's resources and recommended several books and websites for my overall development in ML. During weekends, we used to have video calls with my colleagues, discussing life apart from work. We used to have a lot of fun playing online games like Among Us.

CHALLENGES FACED

One of the biggest challenges that most people face is getting a positive reply to one’s application. I received only two positive replies after reaching out to more than 200 people. Professors generally hesitate to give research opportunities to sophomores, usually due to their lack of genuine interest in the field and inexperience. I would suggest reaching out directly to the professor rather than through the university’s website and having a genuine interest in their research. I would have loved to do research onsite, but unfortunately, I had to do it online. The work experience that you gain working onsite is priceless. Sometimes, due to the difference in time zones, the meeting timings disturbed my sleep cycle a bit.

LEARNINGS

My interest in ML has increased manifolds after I completed my internship. I received a Letter of Appreciation and a Letter of Recommendation from Harvard. These are the best things that a student can hope to get after working on a research project. It even helps if you want to go down the corporate route in the future. Research is always a team effort and the best part about a foreign intern is the long-lasting and multi-dimensional connections you make with like-minded people which will help in one’s future endeavors. The research had a deeper impact on me. It changed my mentality as a student. Being a product of the JEE process, I, like many others, was a result-oriented person. During the internship, I learned to appreciate the process of achieving the results and understand that research is a time-consuming process. The experience instilled in me a sense of patience and discipline.

TIPS FOR JUNIORS

Utilize your first and second year in exploring new ideas and technologies. Resources like research papers and websites like Google DeepMind are really helpful. Gradually try to figure out which fields appeal to you more. Don’t hesitate in reaching out to the professors working in the field of your interest. Always pursue your interests and avoid running after trends. I just want to clarify that high-paying roles exist in every field; if you love your work, then you can definitely make it big. Sometimes, other than emailing, reaching out to people through platforms like LinkedIn can also provide you with opportunities and resources in the field of your interest. It is a misconception amongst the community that top colleges do not give a positive reply. I believe that it’s true that getting a research opportunity at a top college is difficult, but if you have genuine interest and passion in a field then opportunities do come. In the end luck also plays a big role while you apply for a research internship. I would advise everyone to let your voice be heard by constantly reaching out to new people and seeking opportunities.