Interview with Ankur Abhinav, CSE 2019, IIT KGP

Ankur Abhinav graduated from IIT Kharagpur from the Department of Computer Science in 2019. He is currently working as a software engineer in Machine learning team at Mercari, Japan. Ankur has also worked at Adobe Inc. as a research intern. We got an opportunity to have a talk with him regarding software skills which he uses and on what things, a student must focus to get into a good software company.

Let’s have a look at it –

What skills do you use while working at Mercari ?

I use basic ML skills like model training and evaluation, creating PoC of a project, system development and API design.

What are the important technical skills, a student must know to secure a good job in software profile?

From an interview perspective, competitive programming is the most important skill. While working at a company, the work depends on the project and team you have been assigned. Depending on the company and also the team inside the company, the work and technical skills required for that can vary a lot. You can easily learn the required skills while working at the company. Mastering every skill at college is almost impossible. Therefore, competitive programming is asked during interviews to test the thinking and implementation skills. Although you will rarely use this while working at the company.

Also, It’s always good to have the experience of working on some projects. Your past projects may not be totally relevant with your work at the company. But it shows how you can contribute to the team you are working with.

How do you see the understanding level of Machine learning or Data Analytics of college students as compared to what is required in the industry ?

In college, when students work on ML related projects, the students think the only thing that is important is modelling, i.e training model on a dataset and evaluating it with an aim to get the best accuracy. This may be good for a research project but while working in a company, you need to work on many other things even if you are in an ML team. Model training is just the tip of an iceberg. You need to work on data cleaning, feature analysis, preparing a refined dataset, deploying your model after training, cron job, cloud, system development, etc. The latency of the model is a very important factor.

How much is the role of ML skills vs Full-stack skills, for cracking the job at Mercari ?

I am not sure. I guess both are equally important for the ML teams while only development skills are important for the non ML teams.

Do companies look for your ML skills more than your software development skills ?

I think this depends on specific companies. If the prospective work is research based then they may look for ML related skills and background but for the software developer role which forms the majority in the tech companies they will certainly look at the development skills. Even if you are being evaluated for an ML team, software development skills are required and very important for your work.

Ankur loves reading books and watching movies. He is a fitness freak and never misses his gym :p. His other fields of interests are Finance, Trading & Investments. Recently, he has started writing blogs on topics such as finance, his experience on software development etc. You can follow these on //ankurabhinav.com/. We thank Ankur Abhinav for giving us few minutes from his busy schedule.