The Placement Journeys
The Placement Journeys — Episode 3
Placement Journeys through the eyes of the ex chair of the ACM Student Chapter at ASE-B
IMPORTANT
Due to the complex situation the world is in right now, I know there has been a significant lag between this and the previous episode. But, I hope you all would not mind that delay because, we have a whole lot planned for you in the coming episodes of TPJ!
Hello! This is Ramshankar and this is the official release of episode 3 of TPJ. In this episode, I bring to you the interview of Arvind Sudheer, the ex-ACM Chair at ASE-B. If you happen to be someone who knows your CSE seniors, this guy is too hard to miss given his absolutely nice nature and inherent excitement when he meets new people! And if any of you happen to have heard of the ACM SIGs (of course you would have), this is the person who first conceived the idea of introducing the global concept of SIGs(Self Interest Groups) into the midst of out-of-class learning opportunities at ASE-B. Curently, Arvind is interning at Caterpillar where he works as a Data Science Intern (Marketing and Branding Wing). He has also bagged a full time offer at Mu Sigma, a company known for the in-depth exposure it provides its employees with. He will be joining here at the end of his internship as a Decision Scientist.
NOTE
Since I am unable to conduct live in-person interviews due to the lockdown, I shall tweak the format of TPJ a little and post Q&A transcripts of my telephonic conversations with the guests. Hope that works just fine.
The Interview
RY : What were the fears you had before placement season?
Arvind : Well, I had a few of them. As a CSE student, we have a broad range of topics to cover starting from traditional ones such as data structures, networking and software engineering among others, to currently trending ideas of machine learning, AI, data science etc. And each company is interested in a separate topic, so it becomes a bit difficult when we need to learn all these concepts together and be able to answer questions. So, the overwhelming amount of content to prepare was the most important fear factor. Another fear was my ability to convince the interviewer of my knowledge. Knowing is one thing, but being able to express it in the right sense when it matters is another. I was also really concerned about the preliminary round of selection. This is the most important round as literally a huge deal of whatever we are asked in the later stages(if we qualify) depends either directly or indirectly on the first round. So, clearing this was very important.
RY : Arvind, having talked about the preliminary round, I wanted to know if you were referring to the online coding test. And if you were, how were your preparations leading upto the first round in any interview?
Arvind : Yes. I was referring to the coding round when I said the first round. Most companies have an online coding test as their preliminary screening round. As far as my prep was concerned leading upto the rounds, I did practice a lot on Geeks for Geeks. Went through their algorithms, code snippets, tried coding them on my own etc. etc. Basically, it was a lot of practice and I guess there is really no other way to do it than practice.
RY : In your view, how do you think a student can improve his or her abilities to present their knowledge and answers better in an interview?
Arvind : What worked for me was mock sessions. I used to explain the things I knew to my friends, family and most importantly to myself. These mock sessions help you infuse your learning in your everyday life and gradually it becomes a part of you. The ultimate aim to a mock session should ideally be to ensure that another person who hears from you must understand to a decent extent such that they can reproduce your work. If you can do this, there is a pretty good chance that you know what you are talking about.
RY : A very common question from students who have approached me in the past has been, “Which academic courses are important?”. What answer would you like to give to that?
Arvind : (straightfaced) ALL. While certain courses in the 3rd and 4th sem like data structures, operating systems are really important from a foundational perspective, I feel all courses we study are fairly important. Because, as I said before, every company has a different requirement. So, if you end up looking at one course as not important, you might be shutting yourself away from one company. And the 1st and 2nd year students should try and explore as much as they can with all courses, even if its a subject they find very easy. Because, you won’t get this time of chill back (pensive look)
RY : How was the interview experience from the perspective of how nice the interviewers were?
Arvind : All the interviewers were very warm and nice. Frankly, the only person in the room who was not relaxed was me. I guess when we walk into that room, it’s natural tendency to get stiff (chuckles). But yes, the interviewers were really nice. During the arista interview, the interviewer even handed over a biscuit to me! The honeywell interview and the one with mu sigma too were conducted by some really nice people. Even in areas were I was kind of stumbling, they were pretty encouraging.
In the next segment of the interview, we had a few questions that were chosen from the question pool provided by members of the ACM Data Science SIG at ASE-B.
NOTE : The most frequently asked questions were segregated and chosen for this segment.
R.Y : What tools and concepts were the centre of your focus while working on data science projects at college?
Arvind : As far as tools were concerned, I coded almost all my data science related projects in python. I used hosted jupyter notebook services such as kaggle and colab extensively. But, I really feel the tools you use is immaterial to a large extent. In my interview they were more interested in knowing about the way in which I handled my problems and how I was trying to solve them.
R.Y : In your view, what would be the best way to approach a project?
Arvind : Stick to a research methodology, have a neat structure and use projects as a way of learning foundational concepts in depth. Languages and libraries keep evolving over time, so perfecting yourself in one is not as important as perfecting your concepts.
R.Y : How was the interview experience at Caterpillar? They are predominantly a Mechanical Engineering company, so how were the questions like?
Arvind : The questions were pretty much similar to what were asked at to other companies. They were questions about my programming ability, difference between R and Python, the packages I use etc. They were very interested in knowing if I knew what was happening inside a package. The questions that were not similar to what were asked in other CSE companies mainly were based on data scientist relevant skills such as intuition as a data scientist, the ability to see the business outlook of a problem and communication ability. Also, having a background in CSE really helps as writing code might be possible for any student, irrespective of their branch. But, writing efficient code and following good coding principles usually comes more naturally to those trained as CSE students than other branches.
R.Y : What are the kind of projects you have worked on at college?
Arvind : A couple of important data-related projects that I worked on were a project on “Detecting criminal activity hotspots” and “Identifying defects on external surfaces of buildings”.
R.Y : Since you have begun working, what do you think is the difference between academic data science projects and company data science projects?
Arvind : Good question. The differences are most prominent in :
- The Data : It’s more sensitive in the company, we are asked to derive business understanding of it, it’s not as structured and clean as the data we pick up for academic projects
- Preprocessing : In company projects, this can take several days. In academic projects, we kind of wrap this up within a few hours.
- The Modelling : In companies, we will have to create our own models based on our needs. So, using ready-made packages won’t always help. The stakes are much higher in companies, so the models have to be better suited for the best results pertaining to our domain.
Most Important Tips Stressed On
- Understand the Inner Workings of the Package
CSE as a branch, is built on the foundations of making things more efficient and simple. But, its highly important to understand how certain things work and looking at the packages we use as black box abstractions are not really going to help once we step out of college. - Don’t Follow the Hype
The hype around data science is real and glorified a little too much by the media. Don’t charge based on the hype, make sure that this is what you really want to do. - Learn To Code
Coding is the most important part of getting a data science job. So, its required that you are thorough when it comes to writing good, efficient and reproducible code. - Be Strong with Foundations
Just like any other job, understanding the basics is very important for data science. Knowing the math behind certain algorithms will help you optimize your models for better performance.
Important Links
- Geeks for Geeks
- Detecting Criminal Activity Hotspots [Project Repository]
That’s the end of episode 3. Hope you liked the content and the new format. For those of you who have missed the first 2 episodes, you can find them here.
Happy Reading! 😄