Welcome back to Kaggle Grandmaster Series!
Today we are thrilled to be joined by Tarun Paparaju to share his journey with us.
Tarun is a Kaggle Notebooks Grandmaster. He ranks 25th in the category and has 18 Gold Medals to his name. Also, he is an Expert in the Kaggle Competitions Category.
Currently, Tarun is a Computer Science undergraduate student at the University of Massachusetts. Despite being this young his notebooks are highly respected in the Kaggle Community.
Note: This is a video interview. We have provided the synopsis below in text. Since this interview was recorded in January, some Kaggle ranking figures differ from what we have mentioned above.
Tarun’s Data Science Journey
The following are some highlights from Tarun’s interview. For detailed answers, watch the video.
Analytics Vidhya(AV): What sparked your interest in the field of ML and Research?
Tarun Paparaju(TP): When I was in 7th or 8th grade I was very much into mathematics other than any subject. I would often buy books from outside school and study them. So I was a mathematician before anything else.
The 3-4 years back my father, who is also a Data Scientist introduced me to Kaggle and that was the beginning.
AV: What was your motivation behind starting to compete in Kaggle?
TP: He never had a fixed goal while participating in Kaggle. He was in love with the process and was trying to learn each and every aspect of data science, from data collection to model building. There was no goal to reach the top. He was fine even at the bottom 10% of a Kaggle Competition as long as he was able to learn.
He spent his spare time searching for random datasets to practice on.
AV: Since you are still a student, a Computer Science Freshman at the University of Massachusetts, how has your education at school and college helped you in your Kaggle competitions or any other Data Science hackathons?
TP: The general route that a normal data scientist takes is – From understanding computer science core to understanding Machine Learning Core Algorithms and, then understanding the maths. He started his journey the other way round.
Tarun feels that the gaps his journey has will be filled by his graduation program.
AV: What resources helped you in Studying Machine Learning?
TP: Tarun suggests these 3 resources-
Kaggle is a great resource not only to practice on random data sets but also to learn from the discussions.
Open Source Contributions and Github
This is one of the best ways to contribute to open-source projects and get your work checked and optimized by multiple people.
This is one of the best platforms for free resources in Mathematics which serve as the core towards data science.
AV: You are a Kaggle Notebooks Grandmaster and you achieved this title at the age of 17 years and you currently Rank 19th. This is something really amazing? So what were the challenges you faced as a 16-17 y/o during this journey and how did you overcome them.
TR: The major challenge Tarun faced was the time management pressure brought by his academic schedule. With weekdays in school and weekends in coaching centers, it was really difficult to find time for Kaggle.
What he did was he put all the academics and social expectations put on him behind and kept his hobby at the forefront. He made sure he prioritized Kaggle over anything.
AV: What is your procedure for creating a good notebook after selecting a dataset? Is there a checklist of must-do tasks to always perform?
TR: In the first 2 years, even after making any kernels no one was paying heed to it. It was getting too frustrating for him as he was putting all his efforts into it.
Then one day, he posted about his frustration on The Kaggle forum, and to his luck, Kaggle Legend Andrey Lukyanenko replied to him. He said the notebooks should be unique and has to solve problems in a way that the reader wants and not what the content creator perceives. That one reply changed his perspective and he started making better kernels.
AV: Since you are an NLP researcher as well, could you name some lesser-known NLP Frameworks which you feel everyone should know?
TP: Tarun mentions 2 libraries that are lesser-known, but has huge utilities-
This is one of the best NLP libraries for multiple languages. Does the work of language detection superbly.
Pattern offers a neat API for classic NLP utilities.
AV: How do you keep yourself updated with all the rapid advancements being made in the field of deep learning since it is a broad area in itself?
TP: Tarun is pretty motivated about following the right people on Twitter. 99% of the time he says, that majority of the things he wants to know about comes from Twitter only. The benefits of having tightly knit social media are clearly visible if you follow the right people and institutions.
In fact, some of his researches are also inspired by the papers and developments he gets to know on Twitter.
His thoughts and words are enough to get anyone to begin and stay focused on their data science journey. I hope this edition of the Kaggle Grandmaster Series with Tarun adds value to your data science journey.
This is the 22nd interview in the Kaggle Grandmasters Series. You can read the previous few in the following links-
- Kaggle Grandmaster Series – Exclusive Interview with Kaggle Notebooks Grandmaster Kostiantyn Isaienkov (#Rank 8)
- Kaggle Grandmaster Series – Exclusive Interview with Kaggle Quadruple/4x Grandmaster Rohan Rao
What did you learn from this interview? Are there other data science leaders you would want us to interview for the Kaggle Grandmaster Series? Let me know in the comments section below!
You can also read this article on our Mobile APP