Biography

I am a Computer Vision Scientist (Ph.D.) who builds deep learning systems that run faster without losing accuracy — and I am now looking to bring that work from the lab into industry.

My core research cuts training compute time for any CNN-based deep learning model by 30–70% with no accuracy loss, using a layer-wise early-stopping method I developed and validated across 6 architectures and 13 datasets (including 10 biomedical imaging benchmarks). A related algorithm I designed for CNN kernel selection outperformed a comparable approach built by a team at Google, improving model generalization with zero added parameters.

Since 2023, I have been an Assistant Professor of Computer Science at Texas A&M University–San Antonio, where I teach and mentor the next generation of engineers. Before my Ph.D., I spent 3 years as a software engineer, including a year leading a 5-person engineering team.

What draws me to computer vision is that the underlying theory does not change with the application. The physics of image formation and sensing, the radiometry and reflectance that govern how a surfaces brightness becomes an image intensity, the geometry of camera calibration and 3D reconstruction, the feature-detection math behind edge and boundary detection, and the math of Kalman/EKF-IMM tracking and multi-sensor fusion are all the same whether the image comes from a defense sensor, a wafer-inspection camera, or an MRI scanner — only the constraints change. I still find myself learning something new every time I sit down with a problem — that constant discovery is what makes this field endlessly interesting to me.

Email:sahmad@tamusa.edu