In the morning, I tend to think of new problems, read research papers. Then, depending on whatever meeting is coming up, I complete relevant work for that meeting. I devote my evenings to development work. Around 7-7:30, I switch off from work. This is when I revise some of my basics. In my work, concepts of linear algebra and probability and statistics are involved. I did these courses in my first year and now I’ve forgotten some aspects of these, so I need to revise them consistently.

Yes, my job has given a direction on what I want to work on. During my PS1, I worked on search enhancement, more towards data mining and machine learning. In 3-2, I took up parallel computing. At Goldman Sachs also, I was working on parallel computing. I wished to work at the amalgamation of parallel computing and machine learning. American Labs has given me the opportunity to do this. There is this field called machine learning systems that lies at the intersection of both, and that’s what I’m working on currently. So, I’d say American Labs has given me a path.

I tried to explore as many fields as possible when I was in college. In my PS1 I had to work on a search engine. That made me think about how we can scale such systems. That made me take up parallel computing. Since I was working on search optimization, I thought of taking up data mining. In 4-1, I took up machine learning and NNFL. I also took up computer networks (CDC), but I didn't develop that much interest , so I did not explore it further. Based on the experience I gained in PS1, courses and internship, I tried to decide what I wanted to do.

There was a time I wanted to go for Masters right after undergrad. But based on my experience in projects and internships, I wasn’t sure which particular field I’d want to do my Masters in. I had just discovered I wanted to put AI and distributed systems together. It was at American Labs that I learnt of this field called machine learning systems. That’s when I finally had some clarity on the field I’d wanna pursue my Masters in. In college, I didn't have a very strong research background. This experience at AILabs has boosted my CV when it comes to Masters Applications. But some people opt for Masters right after their undergrad because it keeps them in the flow of studying and performing research. For some people, if you start earning in a job, you feel like “I am earning now, why should I switch?" In your bachelors if you’re confident of what you wanna do, you can opt for Masters. And then sometimes there are financial issues that play a role in making such decisions.

In Europe, the colleges that I checked had a more research based Masters. I wanted an industry based Masters. The second reason was job prospects. Jobs in Europe have a requirement of European language. But yes, it's easy to get a permanent resident there as compared to the US.

I gave the GRE twice. I gave it in 2018, that was my 4-2. I thought I would prepare well in my last semester but I didn't end up doing so. I prepared in the winter vacations only, I scored 317 and thought I could do better. I took up Byju’s GRE prep which prepped me. Then I scored 323, and that’s the score I applied with.

It gave me exposure to AI. The search engine I made was to be integrated in schools. This comes under the umbrella of AI for Social Good. My PS1 has inspired me to work in this field more. In the future, I’d like to work in a company that works for social good. This internship has also influenced some of my future decisions. The NNFL project I chose was again based on search engines and questions answering algorithms. Recently, at AmEx, we had an idea pitching competition. I proposed something on the line of the work I’d done at my PS1, and the idea got selected.

I got my SN Bose result around the end of February. After that, I started emailing professors and asked them if they were willing to take me. Most of the professors had their slots full. And for those who agreed to take me, they told me that the SN Bose funding of $2000 is less, and I’d have to shed in some money from my pocket as well. That's why I thought it might not be with it. First of all, I didn’t know how the guidance of the professor would be. At that time, I wasn't aware that I could contact MS/PhD students of those professors and find out how he is as a supervisor. Secondly, I was spending money rather than earning money. There was also a job factor, I thought if I could get a PPO at Goldman Sachs, that would be a great backup.

My work at Goldman Sachs was more on the SDE side. I had to handle the backend of a web application using Java. In SDE, you have this concept of Agile Development where your development is divided into sprints of 2-3 weeks. In each sprint, you are given small 2-3 tasks every day. You have standups everyday where you provide updates and share the problems you're facing. There was more structure to the work here. I was directly given a set of issues that were complete in themselves and I had to simply work on them. At AmEx, it is more of a research-like setup. It is difficult to define your task. You are given an open ended problem and aren't sure how much time the research would take. You have to define your own milestones. You don't know what directions the research would take and how it would work out. During my internship, I worked on distributed components. I started exploring research papers to improve a particular component of the distributed system, but I wasn’t getting satisfactory results. So I had to roll back and think again. There is development work at AmEx, but there is also a research part attached to it. In research oriented roles, given a more open ended problem you yourself are leading that particular thing, so that gives you more freedom on how you want to mould it.

My internship at Goldman Sachs was dominantly based on software development. I liked SDE. But it so happened that I didn't get a PPO at Goldman. I sat for placements, where I got into Tesco. So, I had a backup at Tesco. My plan was to apply for MS at that time, but I didn't end up doing so because the placement season went long for me. Then PS2 started, I applied for Nutanix and American Express. I knew that at American Express, I’d be getting work in machine learning but the role that I’d be getting at Nutanix wasn’t that clear. So, I opted for AmEx. I got a PPO here, the selection for which is based on the work you’ve done there, and your profile. Since I’ve enjoyed working at AmEx, I decided to continue with this. Also, I wanted to pursue a Masters, and I thought experience in a research-oriented role would strengthen my application.

During research you do run into roadblocks, and you learn stuff the hard way. I just take a day or two off after that. I talk to my manager: he has done his PhD and PostDoc, so he is able to guide and motivate me when I hit a low. I try starting afresh after I take a break.

No, not particularly. Everything felt in place in a very good way. Probably not taking a project in 2-2 itself is a regret. Doing a project then would have given me a good head start, maybe help me publish a paper as well. This also caused me an issue on getting LoRs because professors give you better LoRs when you have a project under them.

Parallel Computing xD. The course was considered to be one of the toughest courses. I wanted to explore it. It had like 36-40 students, and two people received an A and 5-6 received A-.

Explore as much as you can. Develop a habit to read research papers. Be really active on Kaggle if you want to pursue a Masters or have a role in data science or artificial intelligence.