Eric has been finishing his Master’s in Statistics at Columbia University. Models and statistical analysis will prove useful at his Data Science internship at Johnson & Johnson. He will be sharing his stories and insights through his blog.
Were you surprised to find a professional match with Johnson & Johnson?
Yes, I was completely taken by surprise. It hadn't occurred to me that Johnson and Johnson was looking to hire people with knowledge in statistics, machine learning, or data science. Lucky for me, they started their data science internship program only a few years ago, so I get to be one of the lucky ones to help shape the influence and future direction of data science at the company.
What has been an interesting learning for you while working with data?
In my deep learning course, I've enjoyed seeing firsthand how effective deep neural networks can be at accurately learning features from data. For the course final project, I worked on a team that achieved a classification accuracy of > 99% on the well known MNIST dataset. This is a remarkable result because this level of accuracy rivals human levels of performance. It's really quite fascinating that such a result is achievable despite the fact that the scientific community still doesn't really have a good explanation for how neural networks do what they do.
3 songs that will be on your playlist this summer?
Believe it or not, I listen to a lot of pop country. First 3 songs that played when I hit shuffle: Days Go By (Keith Urban), We Are Tonight (Billy Currington), and American Honey (Lady Antebellum).