Hello everyone, I am Apoorva Kumar, and I recently graduated from the EEE Department of IIT Guwahati. I am currently employed as an Associate Data Scientist at NoBroker.com. While my placement journey may sound unconventional, it surely was a memorable one and will help anyone pursuing the data science profile immensely. I will start with my 3rd year intern followed by why I chose Data Science and what other plans I had, to how I ended up being placed here. Please bear with me 😅.
The Intern which changed my perspective
So I landed a very lucrative internship at a Semiconductor firm in South Korea named Protec Inc. who are one of the lead suppliers for Samsung. I joined a small team which worked on using Machine Learning and Robotics to automate chip designing. My main project was to design a Stereo Camera based Odometry ( a technique to track the path of a moving object ) for a self driving vehicle. It was an interesting blend of Robotic and Machine Learning to build something new and interesting. We started by building a depth image from stereo camera using computer vision techniques to capture features which will be used for understanding how the robot is moving. We implemented multiple techniques to perform the depth sensing task and chose the technique which was best suited for our scenario. Next we performed odometry using novel features extraction technique in ROS from the depth image and concluded their output for use in production.
The work, both in Robotics and Deep Learning, was wondrous and the state-of-the-art implementation for real world scenarios really gave me a kick and further solidified my interest in the field of ML based automation, Robotics and Data Science.
What is Data Science?
I will start by saying that Data Science or DS is essentially not a very defined field. Like Software Dev have Java and C++ Devs separately, Data Scientist profiles also range from Data Analyst to ML/DL Engineers. I was more interested in the latter. I would like to bring to notice here the very important aspect of reading the job profile completely and looking if it really fits you rather than just applying based on by the name of the profile. “Don’t judge a book by its cover” seems utterly viable here.
These roles capture work ranging from basic data management and analysis to applying state-of-the-art techniques to empower real-world products used by millions of users. Sounds pretty heavy right? Well to be honest the fun really overtakes all the anxiety and effort that goes into making and developing such techniques. Anyone who has a knack for learning new things and automating the world around them should try for these profiles.
Why DS or more appropriately How DS?
I will start by saying that Machine Learning and Robotics has always been my passion since I joined college and I have stuck to that with everything I had. After completing my internship in the same two fields I was more than confident about my profile. I had a clear plan that I would only apply for Machine Learning(ML/DL) profiles ( Robotics profiles sadly don’t visit campus) and won’t sway from it under any circumstances. Data Science as I mentioned above is the broader field under which ML/DL lies, and thus began my journey towards being a Data Scientist.
Before the Interview: The Learning and The Tests
Once I had my eyes set on Data Science profiles all my efforts went in that direction. I only applied for Data Science roles and skipped all Software Developer and other roles. I actually ended up applying for only 21 profiles compared to others who had applied for more than 50 profiles and to be honest this fact had me on my toes till the point that I was placed.
Starting with your resume — it will turn out to be either your best friend or your worst enemy during your interviews. While it might not seem very relevant during the test phase, the resume will play a major role as a point of discussion and forming the first impression of you on your interviewer. Try to make a one-page CV with good work experience and only your top 3 or 4 projects. Keep your description precise and highlight the keywords. Don’t try to squish in things just because you did it but see if they are really relevant to the job you want.
Moving on to what I studied. Topics related to Basic Machine Learning, probability and linear algebra were most important with added concepts of Deep Learning if you have time. I had these two beautiful courses named Fundamentals of AI and Advanced Probability & Random Processes which helped me revise all these topics over the course of my 7th Semester. I did all my preparation using these slides and cleared my doubts by just searching the question on Internet or asking friends. I also focused on mastering Python libraries as it’s the preferred language for Data Science profiles.
Coding, including but not limited to, Data-Structures and Algorithms was also a requirement as some companies used to put coding questions in DS profile tests. I never did GeeksForGeeks or InterviewBit though but mostly solved random problems my friends gave me to get a grip on my coding skills. Python makes coding much easier and most companies allow Python in their tests now.
The tests were fairly easy compared to the coding tests of SoftDev profiles as it was either MCQ based or required you to make a model on the go and submit the results on a test dataset. The coding questions were there but they were either easy or most companies didn’t put a lot of weight onto those questions.
Interview Experience
I never specifically targeted this company up to Day 1.1 and decided to apply here because the role seemed very promising. They didn’t have a test, instead they shortlisted candidates based on CV. This was going to be my 2nd interview in this field.
My first interviewer asked me for my introduction followed by a demo of the app and website of the NoBroker.com by him. He asked me some questions like
- Why Data Science?
- How will NoBroker help me in my future career?
He then asked me a really interesting question, which required me to explain one of my projects whose use case or knowledge base will help me add something substantial to the company. He wanted to learn if I had succeeded in doing something which I can directly replicate on joining the company and contribute to them. I explained the work of my 3rd year intern and how it would help the company build something interesting and new.
He was happy with my answer and also helped me shape my answers on the go. He followed by asking some questions related to basic ML and put some scenarios in front of me and asked me to solve them using ML techniques. He ended by asking me some questions on
- SQL and Pandas
- Probability
- Random Variables
Once the interview was done, he told me to wait for the next round which was with the Founder and CTO of the company, over Skype. He asked me questions like
- Why NoBroker?
- Why did I not pursue research and choose to do a job?
- Some puzzles based on Probability and problem solving.
He then told me that he would love to meet me soon in Bangalore! Both the interviewers and the on spot HR were very polite and kept us really at ease. They even heard my request of interviewing one of my friends who was not shortlisted earlier.
Ending Note
Thank you for hanging on till here! I know it may have been a tiring read but be sure to use anything that you captured here to excel in your interviews and tests. While the placements might seems like a test of knowledge, it’s more often a test of confidence, patience and mental strength. Remain confident and be clear on your priorities.
Always keep your friends by your side and be by their side whenever they need you. Believe me, they will be your greatest help during this whole session. Never hesitate to help or ask for help! All the best and Excelsior!