Reading Time: 7 Minutes

image

“You can’t connect the dots looking forward; you can only connect them looking backward. So you have to trust that the dots will somehow connect in your future. You have to trust in something — your gut, destiny, life, karma, whatever.”

In the second edition of Career Series 2020, we have Sumit Jain, an alumnus from the 2017 who graduated with a B. Tech. degree from the Department of Mechanical Engineering, who talks about his experience as a Deep Learning Engineer at Robert Bosch and his further journey which saw him return to IIT Bombay to pursue a masters in data science.

 

1.What were the various options available to you while choosing a career path? How and why did you decide to pursue a career in the core industry of Mechanical Engineering?

To really describe my career, I think it has been short, but happening until now, with a fair bit of academia mixed with two different kinds of work environments.

I graduated in 2017 with a degree in Mechanical Engineering. During my undergrad, I did a couple of internships abroad, found myself building a student satellite, and at the onset, I was pretty clear that I wanted to work in the core engineering sector, a decision fueled by a fair amount of exposure to what it would be like to work in the industry.

Immediately after I graduated, I landed a job at Robert Bosch, where I was promised a mechanical engineering role – until this point my career path was going exactly according to plan. Things quite suddenly took a turn for the strange when, on the job, I was given the role of IT Management Consultant instead. It knocked me into a completely different career path from where I expected to be. 

For the first few months, it was really hard to digest this shift in paradigm. I found myself getting really anxious, questioning where I was, why I wasn’t doing what I was good at, why I wasn’t where I was “meant to be”. But not everything happened for the worse (as I thought of it then); graduating from IIT had its own perks at the job. I was given a client-facing consultant job, a role that otherwise might’ve taken me 4-5 years to get. Despite the obvious dissatisfaction I was feeling, I was still happy that I could travel around and interact with clients. It was a pretty huge responsibility for someone who was just getting started. 

As I neared the end of my first year at Bosch, I was beginning to realize that I wanted to do something different, but in all the time I’d spent with them, my previous dissatisfaction had changed to unlikely interest. So as I pondered upon the question “what next?”, I found myself at a crossroads – to either start looking for mechanical engineering roles or continue working in the IT sector.

I decided to take one more chance with IT. I began to play around with a lot of data, and I realized I really enjoyed the whole process of converting raw information into data, and then use it to generate useful insights. Concurrently, I was doing what any ML enthusiast would do – taking Andrew NG courses, specializations, watching Youtube videos, and so on. 

As my interest grew, I felt that having formal academic backing would solidify my understanding of the field – through my own exploration I already understood the ‘how?’ part of ML, but still didn’t know the underlying ‘why?’ it works.

This is when I, almost impulsively if I may admit, decided to pursue a masters in data science. Mind you, I was already late for application to any foreign university, and my only remaining options were to either apply to an Indian university or wait another year. It wasn’t an easy choice, really. I didn’t want to wait another year, and there were a lot of opinions against pursuing a masters degree from India, especially given that I already had the coveted “IIT tag” . 

Ultimately, I chose to go with my gut and decided to go ahead with a masters back at my alma mater IIT Bombay itself.

Talking about why it was the obvious choice for me, I was familiar with the institute and luckily enough, the program offered by the IEOR department suited my needs perfectly.

In the next two years of post-grad, I learned a lot of basic as well as advanced stuff. I also did a couple of freelancing projects to get a flavor of real-life applications of ML. I think I can say that I had to work really hard, but I learned a lot during my time at IIT. 

Currently, I’m working as an independent technical consultant with various startups where I do a lot of ML/AI stuff. My role is incredibly flexible and it adds a lot of value to my profile. I’ve ended up very far from where I started out, and it may be rare for someone who is working in the core engineering sector, but I do love my job. 

 

2.What was the general process:

  • Interview (number of rounds, questions asked, topics they questioned about in each round, etc.)
  • Tests (resources)

for the companies that you aimed for? 

I also sat for placements at IIT Bombay, both during my undergrad and master’s. There is a lot of “fuss” around the placement process because of course, the stakes are very high. I think one of the major factors relating to this is that it happens on such a large level and most people only do this once, so it is a little unsettling.

Having walked down that aisle twice, I think it still is a really exciting and memorable part of your journey as a student. The process has become more and more streamlined over the years. Let me talk about the most recent one. I interviewed with a couple of firms. 

Day 1.1 – ANZ Bank – The interview started with a brief introduction and then opened up a little to become a more technical discussion. I think I really felt at ease because they restricted themselves to basic questions related to linear algebra, Python, and statistics initially. We also talked about my thesis, which was something that I could really sell, and they found the topic and work interesting as well. 

This was followed up by a guesstimate question. I was able to chalk down the problem to a reasonably good level framework, but had a little trouble estimating some of the numbers. The overall interview went on for around 50 minutes, which I didn’t quite realize. 

The shortlist was published in a couple of hours, with my name on it. From previous experience and discussion with a few seniors who have been working with ANZ, I had a gut feeling that this would be mostly a discussion on what it would be like to go onboard with them. In round 2, we basically had a discussion on what the actual work at ANZ would be like, and at this point I realized that the work would require a lot of statistics, but not much of ML or Neural networks, where my interest actually lay.

Now in a normal scenario, it would feel really stupid to decline an offer on Day 1, Slot 1, but having gone down that road once, I knew I had to say no because it didn’t really align with what I wanted to do. 

Day 1.2 – Honda Japan – I was waiting for my interview with TSMC when I received a call from one of the coordinators that Honda is also interviewing waitlisted candidates. 

It was something that I hadn’t quite anticipated. 

The interview was taken by Japanese recruiters, and barring a few communication issues, it went quite well. The interview was a lot more technical in the initial phase. The interviewer asked me a couple of open-ended problems relevant to their use cases. There were also questions from data structure and statistics which I was able to work out. In the later part of the interview, the focus shifted from technical stuff to working in Japan. Apart from having the relevant technical knowledge, they wanted a fit who would adjust well within the Japanese culture. I think it’s as odd as it may sound, but all that time spent talking to people at Shirucafe in the middle of multiple coffee runs a day really helped me get the first-hand experience of what it would be like to work with them. Overall, it wasn’t really a hard choice to say yes.

 

3.What made you choose this role of Senior Deep Learning Engineer? What exactly is the work in such a role? Also what hard/soft skills are needed, to be good at it?

In an ideal world, I would be required to move to Japan in October this year, after I graduated in June. This would mean a three-month gap, and I didn’t really want to sit idle and wait for it to happen. So I started taking up small projects and built my portfolio to cater to a larger variety of clients.

Currently, I am working with a startup in the insurance sector, as a Deep Learning Engineer. It’s very hard to describe the usual day in the dynamic environment of an early-stage tech startup. To give you a hint, I find myself alternating between software development, product design, recruitment and more on any given day. The learning curve is really steep and you have to be on your toes. What I do think is a must to work in such an environment is a feeling of ownership, being scrupulous and efficient.

 

4.What are the growth opportunities that exist in your industry? Any personal gratification for your work you’d like to let the readers know?

Lots. I think it’s very cliched but working in a field that you love does really give you a lot of freedom and mental peace. Since I am currently working independently, I have complete freedom on the kind of projects I would want to take on. We recently recruited a few candidates from IIT Bombay and I think life has come full circle. I would still categorize myself as a beginner, but with machine learning permeating different industries, I am hoping to get good projects in the time to come. 

 

5.Would you like to give any advice to people sitting for placements this year, along with some closing remarks?

I think the best piece of advice is to be better prepared and believe in yourself. IITs give you a lot of space to explore and find yourself. At the end of the day, it is important to realize that your IIT degree doesn’t prepare you for your first job, it prepares you for the last one. It is okay to fail and falter, it’s important to try out new things. It is okay to change your mind about what you once thought was right for you. Your career makes up a large part of your life and who you are – it makes sense that you listen to others, but do what feels right for you.