My Journey from a SWE at Google to a PhD student

Recently almost everyone who met me, asked me why I left my (hidden words: high paying) job and decided to go for a PhD. I would normally end up giving them a short answer like "Because I wanted to get into academia" or "Because I wanted to focus more on publishing" or something similar. Also a lot of students have asked me about my application process. In this post I will try to answer the following in detail:

• What exactly led me to quit my job and start my PhD ?
• What was my application process like ?

First of all, the ‘Dr.’ title always attracted me since I graduated from BITS or probably even before. I mean, wouldn't someone address me as ‘Dr. Patel’ sound so cool ?  Though this reason is not good enough and was quickly removed from my mind because a lot of PhDs lately don't really insist on adding the 'Dr.' as a prefix. Anyways, based on my GATE score in 2014, IIT Madras invited me to visit the campus and get selected for their 'Direct PhD programme'. However, a few months later I had my Google offer and instead decided to join Google. So what changed ? Why did I not start my PhD then ?

Well, a couple of reasons. First, I didn't know what my PhD research field would be. I liked algorithms and theory of computations and a few topics in maths. But never really felt a strong enough passion towards them to invest another 4-5 years in a university. Second reason was that the Google offer was very attractive and I didn't really see any big reason to not have that experience which could eventually lead to getting a pic with Sundar Pichai !!!

Sometime after joining Google, I watched a session from their internal research conference from the operations research team. Meanwhile I had taken a couple of courses on coursera about discrete optimization and had really enjoyed them. I got really excited about the fact that there is a set of people doing all that work at Google. They had a few 20% projects so I rushed to my manager to seek permission. My manager said NO. Mainly because I had just joined my team. So I threw a tantrum !! (so professional) and after a few months he saw that I am doing fine with my regular work so he allowed me to do a 20%. That minor struggle kind of increased my motivation for working with the operations research team.

Seeing my work, the team lead at operations research team put some good words with his manager and they hired me full time. I moved to Paris. At that time I did have a lot of passion but not enough knowledge to work with them. Almost all of my teammates had finished their PhDs long ago and had been in the operations research field for a long time. So naturally, I got "imposter syndrome" (a term that I learned after joining that team). I started reading a lot of papers and slowly started gaining expertise about the optimization solver I was working on. I was anyway the only person in that team who was working on that solver full time. Other team members had other projects as well. So, within sometime, I became the go to person for the practical queries regarding that solver. My team lead started trusting the experimental ideas I was proposing. Made me feel good. But still, it wasn't enough to push me to go for a PhD.

It was July 2018 when I decided to attend my first external research conference. I travelled to Bordeaux (in France) for ISMP and then to Valencia (Spain) for EURO conferences. I loved the presentations. A lot of them were by PhD students. That is when the fact kicked in that even after spending more than a year in the research product area, I hadn't published a single paper. I had been doing a lot of experiments myself. My manager wasn't really against publishing. So what stopped me ? Nothing. But at the same time nothing really motivated me as well. Even though I was in the research product area, my job role was "Software Engineer" and not "Research Scientist". So my evaluations, ratings and promotion mainly depended on how much performance improvement my solver was getting instead of how many papers I was publishing. So I never focused on conducting experiments that could lead to a publication or publishing any of my previous work that improved the solver. Interacting with the researchers in that conference made me really think about going for a PhD.

Well, I should also factor the part that I got out of a very long relationship a couple of months ago so I was emotionally not very stable. But I guess it was mostly the second dose of imposter syndrome and the fact that no one really knew me or any of my work, that started the chain of thoughts for going for PhD. A lot of the presentations in that conference were very math heavy. I also kind of felt that my foundations in optimization are not strong because I didn't formally study those courses. I was wrong about that but I could realise that after a year into my PhD ! Anyways, as I attended more conferences, it became clear to me that going for a PhD would force me to publish papers and strengthen my academic foundations. Unless I start publishing, there is no way I could get any recognition in the operations research community.

I talked about my desire to go for a PhD to my manager and he was surprisingly very supportive of that. He suggested that I should move to the USA and work from the Cambridge office (where he was working from). That would make my application process to universities in the USA a bit easier. Also MIT was right in front of the Google office so I can attend talks and classes if I want to. I would get to attend a conference in the USA and meet faculties which could help with my application process. And it did. The PhD offers I received were mainly because I had face to face interactions with the advisors.

I started preparing my application for my PhD in 2019 after I moved to Cambridge. I cleared the necessary exams (GRE and TOEFL) with a decent score. I started writing my 'Statement of Purpose (SOP)' document. Given that I had worked on solvers for a long time, my SOP was standing out. But my English was bad. I showed the SOP to my manager and teammates, and they started suggesting edits. I ended up rephrasing almost every sentence from my first draft ! My grades in BITS were good. I had a very high CGPA. That always helps. Finally, the last part was referrals. Although I had impressed my BITS faculties during my undergraduation, I wasn't sure if their referrals would help me because they were not primarily conducting their research in operations research and my undergraduate thesis had nothing to do with discrete optimization. But I needed referrals from academia and industry. My team lead at Google and my manager wrote some referrals and I contacted my BITS faculties for the others. I was in touch with them so they knew a little bit about what I was working on. All those good referrals, specially from my team lead who is very well known in the community, lead me to getting my PhD offers from Monash University (Australia) and Polytechnique Montreal (Canada) to work with the advisors I really wanted to work with. They had been working on the projects that matched my work at Google. I voluntarily stopped my application process at CMU because of those offers. Mainly because a faculty member who was evaluating my profile there told me that they really liked my profile and I was at a stage in the application process where if I get an offer and reject it, that slot would not be filled by another student.

However, around the same time, Covid happened and everything got closed. I couldn't move to Melbourne and Monash University kept delaying my starting date. I wasn’t sure if I would ever get to start my PhD with them since my offer was about to expire and I would have to go through the application process again. Meanwhile, moving to Canada was easier because I was in the USA and the borders between two countries were open for students. So I finally started my PhD in Polytechnique Montreal in September 2020.

Finally, here is my advice if you are planning to apply for PhD programs. The work experience on your resume is always valued no matter where you submit it. In my case, since my job was in the research division, my work experience directly helped me get offers from universities. The exams (GRE and TOEFL) don’t really play a big role for PhD admission decisions. As my teammates explained to me, they are used for some basic filters. So any decent score would do. Don’t try to optimize it. Instead focus more on building a better connection to get strong referrals. The best way to do so is by attending research conferences. Getting clarity on why you want to go for PhD and knowing what kind of research you want to do should be helpful for your SOP.

Thanks for reading ! More posts to come soon….