Efficient Deep Learning / Signal Processing / Hardware-Aware AI
I'm a Ph.D. candidate in Electrical and Computer Engineering at the University of Illinois Chicago. I build neural networks that are efficient by construction: instead of compressing a large model after the fact, I replace expensive learned operations with fast signal-processing transforms (Walsh–Hadamard, DCT, wavelets).
This has led to multiplication-reduced convolutions, decorrelated attention for Vision Transformers, and transform-fused state-space (Mamba) models. I take these designs through to real-time and FPGA deployment, and test them on prostate MRI, neural-oscillation tracking, and wildfire remote sensing. Highlights include a Walsh–Hadamard fire-detection model with 6.6× fewer FLOPs than prior methods, a 283 µs FPGA neural-phase estimator (IEEE TBME), and an open wildfire dataset with 20K+ downloads on Kaggle.
My work includes 8 peer-reviewed papers, among them IJCAI 2026 and IEEE TBME. I defend my dissertation in January 2027 (degree conferred May 2027) and am available from January 2027 for Research Scientist roles in efficient and on-device ML. CV · Email
Our open-source wildfire dataset on Kaggle has reached approximately 20,000 downloads. The accompanying paper, Large-Scale Open-Source Image and Video Dataset for Robust Wildfire Detection and Classification, was accepted to the IEEE ICIP 2026 HydroImaging Workshop.
View the dataset on KaggleAdvisors
I am grateful to work with advisors whose research has shaped my direction in efficient AI, signal processing, computer vision, and medical imaging.
Explore
Collaborators
University of Illinois Chicago
Northwestern University
Bagci Lab
University of Chicago
Fermilab
USDA