Currently working towards a published paper that proposes a new Neighbor2Neighbor method to de-noise brain PET scans via an unsupervised convolutional neural network. Using the PyTorch library to develop the unsupervised model.Successfully developed a binary classification model that distinguishes between noisy and clean medical images, achieving a test set F1 score of 0.946 and accuracy of 0.944First undergraduate to be offered a position in this graduate lab

Developing a domain transfer CycleGAN to convert noisy echocardiogram to denoised echocardiograms using clean corresponding MRIs as ground truth. The denoised echocardiograms exhibit learned structural MRI features.Designed a 3D printable CAD model of the heart from a CT scan. Model is hollow with open valves and accurate wall thickness, suitable for cardiac device testing.

Responsible for customer service, supervising a team of cashiers, closing the store, preparing for the next morning opening, and tracking daily monetary company balances. Flexible worker who assisted with other departments as needed.