Om Duggineni is a first-year student at the University of Maryland, College Park, and thoroughly enjoys everything science and computer science. He is passionate about the possibilities at the intersection of biology, computer science, and engineering. He is conducting research whenever possible and pursuing researchopportunities and mentorship. When he is not into science, he loves to bike, take photos with his mirrorless camera, read, ǝpoɔıun ʞɔɐɥ, sing, and vacation with family. He also likes listening to podcasts - his favorite podcast is Hidden Brain.
If your web application runs untrusted user code, use scriptybox! scriptybox leverages browser technologies in order to limit the script's access to most web APIs, preventing it from interacting with the code of your web application except in specific, well-defined ways.
Through high school, I've done a variety of experiements on protein folding.
In 10th grade, I used a LSTM-based neural network to predict the secondary structure of a protein based on it's amino acid sequence. I was able to achieve about 80% accuracy on the validation set.
In 11th grade, I took this a step further by using a related technique to predict how mutations to the Sars-CoV-2 virus would affect the virus's ability to avoid human antibodies against Sars-CoV-2. I was able to successfully predict future variants of the virus that may be better able to avoid the human immune system.
OtoFind is a smartphone-based otoscope which leverages the power of the smartphone’s inbuilt camera and the power of offline machine learning to quickly and accurately diagnose Otitis Media. Our prototype was capable of achieving 84% accuracy when diagnosing ear infections. This is an inexpensive solution that provides an early, accurate diagnosis to patients with Otitis Media and other ear infections.
AquaBot is an inexpensive, automated stream-monitoring robot built to measure chemical characteristics in waterways such as dissolved oxygen, pH, temperature, and turbidity to identify polluted areas in streams. Our prototype was able to successfully measure temperature and turbidity data in a local lake and transmit the results to a nearby phone via Bluetooth for logging. With AquaBot, we planned to disseminate water quality information in real time to enable conservation action and increase public awareness about stream health using robotics.