INTERVIEW

I had just graduated from college and I was waiting for graduate school to begin. In the meantime, I wanted to apply my Statistics & Mathematics majors into real world practice, and Sumitomo Electric offered exactly what I was looking for. Not only did the technical (machine learning) job descriptions catch my interest, SEI was also internationally famous for its R&D activities. I was sure that this internship would become a great learning experience for me and help me take one step closer to my ultimate goal of becoming a competent data scientist. Therefore, I immediately applied for this internship.

My project was about predicting future traffic commuting times using big datasets that contain past traffic commuting times. Not only did my mentor allow me to use my favorite statistical software, called RStudio, my team also created an Amazon Web Services (AWS) account for me so that I could approach my project as efficiently as possible. Throughout the internship, I felt my team’s keen interest in the work I was doing which motivated me to work diligently. After training various models with different kernel functions and parameter values, I was able to produce several models that can reliably predict future traffic commuting times. As the location and dates of the month changed, distinct kernels and specific parameter values were necessary to accurately predict future commuting times. Through this project, my coding skills have improved and under my mentor’s guidance, I learnt novel ways to evaluate my models as well. Moreover, since I wanted to practice my Japanese skill I also tried to communicate in Japanese to my mentor as much as possible. Although it was challenging to explain all the technical parts in Japanese, it was a really good opportunity to practice for me.

I had heard about Time Series data before, but I had never actually conducted any sort of statistical analysis on such a dataset prior to this internship experience. I learned about interesting imputation methods that worked particularly well on Time Series data and distinctive patterns that were unique to Times Series visualizations. It was definitely more difficult to manipulate Time Series data than most of the data that I had to explore before. After completing this internship however, I have a much clearer idea on how to efficiently and effectively approach Time Series data in the future.

The most valuable takeaway from this internship was that I got to experience what it feels like to be an R&D engineer or data scientist. I thoroughly enjoyed solving various problems that I encountered here and thinking of how to approach the new tasks and challenges I faced daily. I learned a lot. More importantly, this experience helped me further set my mind on becoming a data scientist.

If the job description matches what you are looking for, I strongly recommend you to apply for this internship. I am sure it will turn out be a great chance for you to expand knowledge in your interest area and to explore the beautiful Japanese culture at the same time! Even if you do not know anything about Japan, it is totally fine. Your co-workers, fellow and past interns who are now employees will make sure you enjoy your time there!