Rajat Ranjan, a data scientist at TheMathCompany, has built serious street cred in data science and analytics in a short span. The former Infosian is proficient in machine learning, modelling and visualisation.

“I love interacting with data, creating models that suit business needs, and participating in ML hackathons to learn new technologies and grow professionally. In addition, I like to indulge in art, photography and creative undertakings in my free time,” said Rajat Ranjan.

In an exclusive interview with Analytics India Magazine, Rajat spoke about his journey to become a Machinehack Grandmaster.

AIM: How did your fascination with algorithms begin?

Rajat Ranjan: Coming from a CS background, algorithms and coding were my bread and butter– everything from C, C++ to Java. I did my majors in forecasting using Matlab and R around 2015-16 without any idea of their future potential. Things started getting interesting from Infosys training when I started programming using Python, and I wanted to excel at it.

After that, I was obsessed. So I started solving questions in Hackerrank, Hackerearth, and other competitive coding platforms to learn and grow. My favourites were Project Euler in Hackerrank and Code battles in Hackerearth. After that, I started solving questions for machine learning, and the rest is history.

AIM: What were the initial challenges, and how did you address them?

Rajat Ranjan: Initially, the biggest challenge was understanding the importance of programming, solving complex, competitive coding questions and participating in hackathons. And the only way I could overcome this challenge was to code and solve – till the end. And I had many hiccups along the way. I had to rely on the community to understand how it’s done. The limited data science resources were also a challenge. Acing hackathons required a lot of smart work.

AIM: What about coding excites you the most?

Rajat Ranjan: I have been hooked on solving coding challenges since my training days at Infosys. The code battles and clearing all the test cases for a question gave me a lot of satisfaction. Since then, I have learned different languages, like Python, SQL, Java and frameworks like JavaScript, React etc. Moreover, I took up a lot of personal projects that helped me understand more about architecture and deployment.

AIM: How do you get into the zone?

Rajat Ranjan: I automatically get into the zone when the problem excites me. I can do it from anywhere with some soft music and a bottle of water. Though it may seem chaotic at the start, you will find the light at the end of the tunnel. The point is to have fun while in the zone.

AIM: What does your ML tool stack look like?

Rajat Ranjan: That depends on the problem statement you are trying to solve. Kaggle gives everything you need, and I don’t have to rely on my PC’s hardware and computing power.

I use sklearn for all the basic algorithms and preprocessing tasks. Then, I use tree-based algorithms like LGBM and Catboost. I use SHAP to understand black-box models. Finally, I rely on feature engineering for domain understanding and use Optuna for optimisation.

AIM: What’s your advice for aspirants preparing for their first hackathon?

Rajat Ranjan: I guess the prerequisites were pretty simple for me. But, of course, it is always Python at the start. But then, for any ML hackathon, it comes down to good domain understanding. Then, dive deep into the sklearn package for error metrics, model algorithms, cross-validation etc. Most importantly, know how to understand data, train and validate.

AIM: What’s your biggest pet peeve about hackathons?

Rajat Ranjan: I find data leakage and plagiarism the most annoying part of any hackathon from an organising standpoint. Personally, it’s always self-criticism like, “it was this simple. Oh, I could have done that/this”. It bothers you a lot when you have given too much time and resources. And lastly, when a few decimal points bring you down to the fourth position.

AIM: What’s the worst experience you’ve had as a coder?

Rajat Ranjan: As a coder/developer, it’s always when things don’t work out the way you wanted, and sometimes that adds a backlog that reflects on future tasks. But a simple way out is to outdo yourself and learn from your mistakes.

AIM: What drew you to MachineHack? How has your journey been so far?

Rajat Ranjan:  I guess I have been with MachineHack since its inception. I have seen how it has excelled in terms of website, community and hackathons in recent times, and it positively resonates with folks like me. I have participated in many hackathons on different platforms. MachineHack has exceeded itself in terms of the level and community. For the budding data scientist, it’s surely the wise genie from whom anyone can learn and grow.

AIM: What was your first MachineHack competition like?

Rajat Ranjan: My first completion was, I think, “Cats & dogs”, “Predict beer prices”, etc., which I can’t find now on the website, though you can easily find me in the Top 10 in other leaderboards. The first few MachineHack hackathons were a blast.

AIM: How did it feel when you became a Machinehack Grandmaster?

Rajat Ranjan: Titles like this gave me a sense of achievement; it’s like earning a medal that gives you a lot of pride, and you know the kind of hard work and sacrifice you have put in to achieve it.

AIM: Tips to ace MachineHack.

Rajat Ranjan:  I guess it’s about consistency for me. The tech is moving way fast, so you have to be ahead of the curve while solving problems. At first, it will be hard and chaotic, but soon it will be an easy ride. Also, it’s important to learn from different approaches and problem solutions.