Predicting B-Cell linear epitopes in the SARS-CoV-2 Spike protein modeled as a classification model. Random Forest Classifier is used for epitope classification along with feature extraction methods such as n-grams on the protein sequence, in combination with the physicochemical properties of the protein sequences of differing k-mer decompositions of the spike protein. As expected, sequences in the receptor binding motif were detected as potential epitopes.
Modeled COVID-19 infection trends to predict future infection rates using SIR model with social distancing measurements globally. Multiple statistical analysis tools such as a lasso multiple regressor were used on daily time series infection data, social distancing data, and population data to predict infection rates a week in advance.
Created and implemented a college course scheduler software using a user's preferences such as school year, major, credits previously fulfilled, and intensity of desired workload in Java. It was made extensible for Brown university and Cornell University and course were modifiable by the user via the UI. It included a graph data structure, a generic database object that worked with a university specific SQL database generated from data scraping in python.
Research testing facial detection models against human cognitive studies in facial detection.
Studied the facial inverse effect in the classification phase of a facial detection model that used the VGG16 pre-trained model for the facial embedding phase and a SVM classifier for the classification phase.
User Interface and User Experience Projects.
Iterative Design and User Testing of an IOS app for a banking startup called Sable. Generated wireframes using Balsamiq, then prototyped our wireframes in AdobeXD. We performed user testing on our prototype app and performed qualitative analysis on the effectiveness and usability of creating a bank account using our mobile app prototype.
