I would start this with the dilemma I faced initially. I was not interested in SDE based roles. I had not done DSA extensively and had no prior experience with WebD or AppD. Being a EEE department student, I started preparing for Digital Design based roles. But that wasn’t it. I was keen on Robotics and was a core team member of the Aeromodelling club for over a year. So here’s how it went.
Preparation
I didn’t target Data Science or Machine Learning roles at the beginning of my preparation. However, I had a good base in Linear algebra and Calculus that made ML a bit easy to start with. I started with the course everyone does- Machine Learning by Andrew Ng. I completed it and was keen on learning more, so I moved onto the Deep Learning specialization course. I was able to complete all five of them just before November. I read a lot of Towards Data Science articles and also learned a good bit of probability. OpenCV helped a lot in some Image Processing questions. I also did Analytics Vidhya’s topic-wise MCQs, and they were a great help in the tests. And that was it. I felt that my preparation was very little compared to my friends and was a bit worried about what might happen.
Test & Interview
The test contained 4 sections: Deep Learning, Computer Vision, Aptitude and Data-based questions. Most of the questions were very similar and even identical to Analytics Vidhya’s blog. Computer Vision/ Image Processing needed quite a bit of experience, and others were easy. The Interview was the best part. The Interviewer was very friendly, and it felt more of a discussion than an interview. He asked me some questions about my project and some skills they valued. My project was based on a swarm of drones running a DL algorithm to find some objects. My main contribution was in Hardware, but I learnt a lot in DL from the seniors involved in the project. He posed some interesting questions on using the robotics knowledge I had to apply to the application they were experimenting with. He seemed very much convinced and said that we shall meet soon. This was my fourth interview, and I was so happy, thinking that I must be selected. However, when the results came, I had not made it! Distraught, I went on with my courses. Then things took another unexpected turn: I received a call from the HR asking if I was still available. I said yes, upon which I had a short HR round and got selected.
The Intern
The internship was a virtual one. I was assigned to a project where I learnt an Ultrasound Image Processing Algorithm. I can’t go into much more detail than that. The Interviewer was assigned as my mentor, and I had a great time. The work was little compared to others because it was a research internship. We were the ones exploring new possibilities on that front. I had a great time handling significant data points (each data point was around 3GB). It was a great experience for me. However, they did not offer any PPOs.
Preparation tips
As you would have noticed, my preparation was neither the best, nor the most comprehensive. I was lucky enough and had a good base in mathematics. But not everyone has. So I would suggest some things to learn. You need not be optimal, but should be good enough in these:
1. Python (have a strong grip on it)
2. Numpy + Matplotlib + Pandas (as essential as anything)
3. Machine Learning by Andrew NG with python assignments.
4. Kaggle’s Titanic challenge and learning EDA & sklearn.
5. Data Structures and Algorithms(at least till trees) and practice on InterviewBit.
6. Deep learning Specialization by Andrew Ng.
7. All topic-wise questions from Analytics Vidhya.
This should be more than enough for most of the DS/ML companies. In particular for Philips, get some grip on Image Processing/Computer Vision. Learn OpenCV if you can.
To conclude, grab every chance that comes at you. Just find your interests, and you will do just fine. Don’t get disappointed if you are not selected. You will have lots of opportunities.
