Bharat Agarwal (EE1)
Domain: Research
Interned at NTU, Singapore
Mode: Online
Interviewed by: Arnav Dixit
MOTIVATION
I think the process of elimination works very well in understanding where you want to go. Once you start mailing, you will find some professors saying yes while some declining your mail. I got a couple of replies, and they were in the areas of cloud computing, bio-medical and machine learning. I thought that Machine learning in general was the subject I wanted to learn since it extends to multiple fields, especially in electrical engineering. Considering that was something I wanted to learn I thought an internship is the best way to learn a new subject. Moreover I was interested in EEG (Electroencephalography) because it is a unique field which is related to my branch as well. So it presented me with an intersection of electrical engineering and machine learning.
Considering we need to prepare for the third year OCS internships too, I preferred an online intern as offline internships are a huge time investment. So I wanted to balance both a research internship experience as well as prepare for the upcoming OCS Intern Season.
APPLICATION PROCEDURE
The procedure is pretty standard. We go and check the rankings of the colleges and check their faculty pages. After reading their profiles and their research interests, we start mailing the professors based on our interests. You can also send personalized emails to the professors with whom you are greatly interested in working.
Usually people mail 300-400 professors before getting a reply. However, I was lucky in this respect since I got 2 replies after mailing 7 professors. I started mailing in January. I do think that early beginners have better chances of getting in. So I would recommend that people start mailing around late August or early September. I got replies from the professors in January, after which I started reading research papers of the professors, which took some time. At the end of my 4th semester, the professor under whom I am currently working asked me to make a small PPT about EEG so that she could get an idea about my knowledge in this field. Also, she could get a proof of concept about how much I was invested in the project this way. In the week just after the majors, I had to read more about the subject and then I presented the PPT to her during the summer break. Afterwards, it went on pretty smoothly, and we had one meeting every two weeks.
PREPARATION
You generally do not need to prepare beforehand. Before this internship, I had no idea of machine learning. I worked mainly on neural networks and I learned everything after I started working in the internship. Regarding resources, you can always look up Andrew Ng’s courses on Machine Learning. There are also many animated videos on YouTube on convoluted neural networks. Also, you can use ChatGPT to understand specific concepts. I also started reading some research papers. Sometimes, internships can have specific requirements, like my professor asked me to learn about PyTorch, which is a Python library. As far as the theory part is concerned, I reached out to a professor in IITD to understand where I can learn about EEG from. He suggested me some PhD thesis of his students which I later studied.
WORK EXPERIENCE
I would rate my experience as an 8 out of 10. This is simply because there were a couple of things that hindered me from giving my 100%. Firstly, I had to prepare for the OCS Intern season as well. Then, since the internship was online, I did not have access to servers to get my models executed. Also you have to schedule meetings with your professors during an online internship. Luckily, my professor was very responsive. Otherwise, the internship was very insightful and I got to learn about many new things. I worked on RNNs and SVMs. I also got to use a model of a professor from NTU and worked on achieving better results by slightly changing the model and hyperparameters. So that was a very fortunate opportunity.
My expectation from the internship was that I would have to work really hard, and probably I would have to learn too much. However, this was not really the case. It’s not like they will give you some novel thing to discover. Instead, they go slow so that you can catch up. I initially had some doubts regarding my prior knowledge in this field. How does EEG work? How does deep learning work? But I think under the supervision and guidance of a professor, it is always easy to learn as the process continues.
CHALLENGES
There are some difficulties in an online internship. For example, machine learning algorithms take a lot of time to be executed. So we can’t run them on normal laptops and require specialized computers which can do parallel programming. So if this was an offline internship, I would have server access to actually run my code on these computers. Secondly, learning new things can come across as challenging. Professors can guide you, but in the end, you have to do the learning part yourself. The professor will give you a topic and you have to figure out everything about it yourself. These are pretty much all the challenges I faced during the internship.
LEARNINGS AND INSIGHTS
One of my biggest takeaways from this experience is the importance of good resource hunting. When you have zero knowledge about a particular domain, then it is imperative to reach out to the right people and study from the right places which is something that I learnt here. Also, this was a very insightful period for me. Now, I know that I don’t want to pursue a role in deep tech spaces where we build things; instead, I want to pursue an analyst role where I can understand and study the features and find out patterns. So, during the upcoming OCS intern process, I am more willing to take up a product analyst or product management role rather than a deep tech role.
TIPS TO JUNIORS
If you know that you are interested in something, then try to take up an internship in the same domain. When you start applying for an internship, many professors would not reply. So, I think it is not right to compromise on your field of interest just because you are getting positive replies from somebody else. Secondly, use this opportunity to learn. If you don’t have anything to do in the summer, try research internships to explore newer fields that you are interested in. This will also help you in gaining knowledge about different things. So, start applying early, understand your interests and reach out to maximum people.
Comments