I am a PhD student at UC San Diego, where I am working in Statistical Visual Computing Laboratory (SVCL) and advised by Professor Nuno Vasconcelos. My primary research focuses on generative AI and multimodal learning, with a particular interest in understanding the underlying feature spaces of these models. My work aims to better understand the underlying dynamics that govern generative processes and explore methods to improve their efficiency and accuracy.
Additionally, I am also interested in the challenges posed by long-tailed data distributions, specifically examining their effects on scene graph construction and how these distributions influence the reliability of computer vision models in the open world,
If not working you can find me at a movie theatre or on my couch binging a series or an anime!
University of California, San Diego
PhD candidate Electrical and Computer Engineering
Sep 2022 - Now
La Jolla, CA
University of California, San Diego
(Thesis) MS in Electrical and Computer Engineering
Sep 2019 - Jun 2022
La Jolla, CA
International Institute of Information Technology, Hyderabad
(Honors) B.Tech in Electronics and Communication Engineering
Jul 2015 - May 2019
Hyderabad, India
University of California, San Diego
Third year ECE PhD student working with Prof. Nuno Vasconcelos in the Jacobs School of Engineering at University of California, San Diego.
Apr 2020 - Now
La Jolla, CA
Computer Vision Research Intern, Intrinsic (Alphabet)
I worked on improving the efficiency of an existing pose estimation pipeline, achieving a remarkable 6.8x reduction in overall runtime. Additionally, I developed and implemented an Efficient PnP-inspired closed-form solution for multiview camera pose estimation, replacing the previous iterative algorithm. This solution effectively handled both discrete and continuous symmetries of objects, significantly enhancing the robustness and performance of the pipeline.
Jul 2021 - Sep 2021
Palo Alto, CA
International Institute of Information Technology, Hyderabad
I worked on medical imaging, focusing on detecting diseases as part of my honors studies at the Center for Visual Information Technology (CVIT) under the guidance of Prof. Jayanthi Sivaswamy. My contributions included building translation-invariant scattering coefficients, which resulted in approximately 5% improvement across various datasets. Additionally, I developed a Capsule Network-based deep neural network designed to operate directly on 3D OCT volumes for classification, advancing the state of disease detection and analysis in medical imaging.
Jun 2017 - May 2019
Hyderabad, India