Published:
My partner and I retrained the final layers of a 3D convolutional neural network model to identify American sign language words as a final project for CS231N: Deep Learning for Computer Vision. I then spun up a Flask backend and React frontend and built a simple website that allowed users to record a video of themselves signing a word and the model would output its top five guesses.
Published:
The summer of my freshmen year, I worked on summer research project with Pat Burchat to use PyTorch to develop different neural network architecture to emulate the Large Synoptic Survey Telescope (LSST) deblending pipeline in order to better understand systemic bias in weak lensing probes as a result of unrecognized galaxy-galaxy blends.
Published:
As a final project for CS107E: Computer Systems from the Ground Up, my partner and I built a 2D plotting robot using a bare metal raspberry pie, a custom laser cut chassis, and stepper motors purchased off Amazon. I personally did all of the mechanical design and wrote the G-code parser to translate g-code (exported from Inkscape) into steps for our two stepper motors.