I worked on the Quantitative Sinus Endoscopy project, where we augmented the current clinical workflow using computer vision and machine learning. I developed an endonasal endoscopy navigation system, where the endoscope and CT are displayed on the application to aid a surgeon to avoid critical structures and guide the surgery.

As a teaching assistant, I constructed/graded weekly problem sets, held office hours, and taught weekly recitation sessions for the following courses: SP23 Signals & Systems - Convolutions, Fourier Analysis, Signal Processing.F22 Image Processing & Analysis - Image representation, reconstruction, segmentation. Image Fourier analysis, multi-scale analysis. S23 Computer Vision - Classical methods: hough transforms, Fourier, edge detection. Photogrammetry: camera calibration, epipolar/multi-view geometry, reconstruction. Deep learning.

I was involved with designing, implementing, and optimizing computer vision algorithms for an autonomous mosquito dissection robotic system, which will be essential for the development of a malaria vaccine developed by Sanaria Inc. I was tasked with certain quality assurance vision tasks using both classical image processing methods as well as deep learning methods, such as VGGNet and ResNet. I was also responsible for creating an autonomous software pipeline from data acquisition to inference.

I was involved with implementing a visualization framework of a satellite servicing robot capable of teleoperation. This was done with Ogre3D, an object oriented C++ graphics library, as well as Unity. The texture data as well as robot control and movement was streamed from ROS over a network, and Unity was used for visualization.

I was involved with the Quantitative Tomographic Endoscopy Project, where we applied computer vision and photonics to improve tumor detection in colonoscopy.