ConchShell is a wearable ASL translator to help the speech-impaired communicate by converting ASL letters into a voice. I used Python, Google's open-source MediaPipe (largely in C++), and Pytorch for a neural network that was trained on 30 000 samples of training data generated by myself and ASL users I contacted. I used a Raspberry Pi Zero W and camera for testing. My friend and I raised $16 600 in funding. We completed a prototype that operated with 94% accuracy and underwent the DMZ start-up accelerator.

I made use of descriptive statistics and linear and logistic regressions in R to find factors (e.g. resistance to antibiotics, type of pathogen, etc…) that contribute to various patient outcomes in a 300-person patient data set at the National Hospital of Tropical Disease in Hanoi, Vietnam.

I conducted a study, "An Assessment of the Significance of Various Air Pollutants on Public Health on a Global Scale". I assessed the significance of various air pollutants on disability-adjusted life years (attributable to air pollution) and provides policy recommendations based on the economic strength of each country. I gathered data for 5 air pollutants (nitrogen dioxide, sulfur dioxide, carbon dioxide, methane, fluorinated gases) as well as sector-specific data for carbon dioxide from reliable sources such as the WHO and WRI.