Alumni Spotlight: Sumanth Reddy, Data Science Engineer at DraftKings, Inc.
This alumni interview was conducted by Liz Eggleston from Course Report.
Course Report recently caught up with NYC Data Science Academy alumni Sumanth Reddy to discuss how being a poker player relates to data science and his experience searching for a job. Sumanth also shares the interview differences among the most popular data science bootcamps. As one of only two students in his NYC Data Science Academy cohort without a graduate degree, Sumanth proves that at bootcamp, your prior work experience may prove more helpful than advanced studies.
What you were up to before you started at NYC Data Science?
Coming out of college in 2008, I was pre-med before deciding to study physics and pursue a PhD. But, I decided against both for many reasons. I played professional poker for a little while. There were things that were fun about it, but after a couple of years I wanted to transition into something more stable.
I started working at a startup, but I felt pretty stagnant because I was just learning programming and data science but wasn’t able to implement what I’d learned.
I decided I needed to build my portfolio and I started looking up these bootcamps. I liked how short they were because I didn’t feel I needed a year of schooling. I just needed someone experienced in programming to answer my questions and teach me a little bit more about machine learning.
How did your stats experience from college and being a professional poker player factor into this?
I felt like poker was data science at its core. In all my life I have never seen anything like it; it seemed perfect. A lot of jobs are very interested in the fact that I played poker because they see the similarities. So I do believe that poker helped a lot.
Physics, especially quantum mechanics, covered a lot of the core concepts of statistics. It was very similar to data science and the concepts were very important.
I find myself very comfortable thanks to those two things.
Given that you applied to all 3 data science bootcamps, what are some of the differences between Galvanize, Metis and NYC Data Science?
I can’t speak much for Metis because I think their class was full by the time I applied.
At Galvanize, the first thing that I had to do was take a whiteboard coding challenge where I made a function in Python. I also had a question about SQL. It wasn’t extremely complicated. After that, I had another phone call where I had to answer a few statistical questions.
At NYC Data Science, I filled out the application, which contained a couple of coding problems, but I wasn’t under the pressure of a clock or an interviewer. Then I came in for an onsite interview, which just consisted of conversation.
Galvanize and Metis aren’t as lecture-focused as NYC Data Science. NYC Data Science has three hours of lecture and the others only have one. The rest of the time you are supposed to manage on your own. You still have the office hours and everything, but it’s a lot more self-guided. We just always had so much help. I don't know if the other bootcamps had a direct feedback loop with the teachers.
What was your cohort at NYC Data Science like?
We had about 18 students. Compared with the previous and most recent cohorts, we had the most even distribution of girls to guys. I think we were also younger; the oldest person was 35.
I was one of two people out of the 18 that did not have some sort of graduate degree. About half of them have masters and the others PhDs. All very, very accomplished people.
Did you get to work on real world projects?
We started our first project a couple weeks in. We had to do web scraping. We had to pick a website with interesting data and create a question to answer about it. Working alone, we scraped it and created analytics or interesting graphs and visuals.
I’m very sports oriented so I did something about the NBA. The NBA finals had just ended and I wanted to analyze the teams and results. A lot of the star players on the Cavs were injured, and I wanted to see how LeBron James’ numbers fluctuated based on each injury, so I created a visualization.
Did you do a group project?
Yes, we had three more projects. The last two projects were group projects. The first one was a kaggle competition. A kaggle competition is when companies come in and pose problems and allow people to compete. There might be a cash prize, a job opportunity or just an opportunity to practice and learn.
They assigned us groups and we got to pick our own projects and compete. Actually, everyone did pretty well on this. Our group and one other group scored in the top 10%.
We all have our projects on Github.
What was the biggest challenge that you faced during the program?
I thought I was doing pretty well during the first month. Then once we got into machine learning, I started to feel like we were going really fast. I know that other people were already stressed out at that point.
I had to absorb as much as possible and it was all coming a little bit too fast. I didn’t really get to go over things in the detail that I wanted to because I also had to complete my projects at the same time. I didn’t get to catch up on that stuff until the boot camp was over, and I’m still going over it now.
What was the feedback loop like? Were you able to tell instructors that things were going too fast?
Yes, they made it a point to ask us for feedback. They actually gave us incentive to give feedback.On the main page that hosted our lecture slides there was a link that said “Give feedback today.” It could be anonymous or you could post your name. At the end of the week whoever gave the most feedback got a $25 gift certificate. I didn’t get one, but I gave feedback.
On top of that, every couple weeks they had a quick 10-minute forum where everyone could speak freely, because not everyone was giving feedback.
Was there a lot of emphasis on job preparation, interview practice, resume building and things like that?
Yes, we focused on that in the last couple of weeks. They had people come in from a company called 5-Star Resume. They gave us advice about our resume, they helped us touch it up, and they also gave us advice about the interview itself.
The interview tips were very helpful. He showed us how simple social cues can make a huge difference in an interview. For example, don’t bring coffee with you to an interview.
He said, “I know some of this stuff sounds obscure but I know people who have passed every part of the interview and that little thing was what cost them the job.”
Have you been going on interviews since you graduated? Does NYC Data Science have a network of hiring partners?
Yes. I am interviewing and it’s going well.
They helped us set up a lot of interviews through the bootcamp while we were still working on our final projects. They had a bunch of hiring partners come in; most of them were just recruiters but some were higher up. One guy was the principal data scientist and another one, they were building a team. They scheduled a 20 minute interview with each one of us. They also had an employer-specific meetup.
What kinds of positions have you applied for?
I don’t look for anything less than data scientist. I don’t bother with data analyst.
I don’t know that my peers are applying to managerial positions. I think I feel more confident than they do. A lot of them had not even been to a data science interview of any kind before the boot camp, whereas I have been doing it for a few months so I had a good idea of the questions that were coming.
That was another reason why I went to the boot camp, because there were questions in the interviews that I could not answer. I needed to know how to answer them perfectly, and now I do.
Looking back on it, could you have done this on your own without spending the money or were there intangible things that you couldn’t have gotten?
Yes, if you’re smart enough and you have the time and the commitment, anybody can get a graduate degree in computer science just at home doing work on their own.
I can say that, for me, 100% it was worth it – but I will say the boot camp is also about the work you’re putting in and the goals that you have. I really wanted to sponge off of the people around me – not just the teachers, but also all these PhD students around me. It was the most intelligent group of people I’ve been around in such a small space before. I’ve never had such an amazing opportunity.
If you’re doing it on your own it may take you three hours, an entire day to figure out what the problem was because you had never seen it before. But someone with experience will tell you what happened and you’ll figure it out in five minutes.
It’s really about the cost of time; how valuable time is to you. If you have all the time in the world and you’ve got no money, sure, go learn it on your own. But $16,000 for three months of infinite office hours and teaching you from the ground up, absolutely worth it.