Projects
Digital Automotive Art
Designed Python applications to create photomosaic car posters using video game car renderings. Automated screenshots to get 10,000+ images of cars with PyAutoGUI and OpenCV, classified images with Google Cloud Vision and Scikit-Learn, and composited the final 36K pixel image on Azure cloud machines due to high memory demands.
​
Fantasy Hockey Cheat Sheet
Developed a Python tool with Pandas and BeautifulSoup for hockey data scraping and analytics. Visualized team strengths, predicted results, and assessed trades. Resulted in consistent top results for 5 seasons.
Research @ RPI IDEA Lab
Developed a hockey video analytics algorithm to track player locations from frame to frame using SIFT and RANSAC which addressed the moving-object moving-background problem for broadcast video.
1st Place Winner in Natural Language Processing IBM Competition @ RPI
Programmed an AI street vendor using IBM Watson Natural Language Understanding that negotiated commodity prices and competed for sales. Utilizing sentiment analysis and information sourced from the web, the AI dynamically adjusted prices, persuasively promoted goods, and highlighted competitor shortcomings.
​
Project was completed in collaboration with my roommate, source here.