About

I am a Principal Engineer in Computer Vision at MathWorks, working on geometric scene understanding, 3D reconstruction, and scalable visual perception systems. My research focuses on learning efficient representations of real-world 3D environments from visual data, with an emphasis on large-scale and resource-constrained settings. My work has contributed to both research publications and product capabilities in MATLAB and Simulink for 3D vision, LiDAR perception, and reconstruction workflows. I collaborate closely with academic researchers on related problems.

My early work explored contextual cues in visual recognition, including temporal continuity in videos, geometric priors for 3D shapes, unsupervised domain adaptation, object detection, fine-grained recognition, and face recognition.

My MS/PhD was at the UMass Computer Vision Lab, advised by Prof. Erik Learned-Miller, and worked with Profs. Subhransu Maji and Liangliang Cao. Before MathWorks, I was an Applied Scientist at AWS AI Labs in Pasadena, working on visual-language models for document understanding. I have an Erdős number of 4 (Aruni → Subhransu Maji → Jitendra Malik → Fan Chung → Paul Erdős).

Please see my Publications page for more details on my research.

CV (2020): [CV].