My passion for data science led me to pursue multiple data science projects, as well as an internship at CureMetrix Inc.
Machine learning on mammograms
During my internship at CureMetrix Inc., I worked on classification algorithms to identify the presence of benign and malignent structures in mammograms. My work primarily involved designing, training, and testing convolutional neural networks in Caffe, as well as cleaning and organizing our image data.
League of Legends champion recommendations
I really enjoy the game League of Legends and, at some point, I began to wonder, "If I enjoy playing Quinn in the top lane, what other champions might I enjoy?" Using Riot Games' API, I collected player data from around 140,000 players and identified correlations in the champions that they choose to play in ranked matches. I used these correlations to build a recommendation system and detailed the approach in this blog post.
Home Depot product search relevance
In this Kaggle competition, we were given a search term and the products that Home Depot's search algorithm showed the customer. Our goal was to predict the relevance of these products given the product descriptions, product attributes, and the search term. Using the tools of Natural Language Processing and a bit of creativity, we engineering predictive features and trained several regression algorithms to predict the relvance of each product. Our final algorithm, a gradient boosted regressor, was in the top 7% of the competition.