Yan Kang

Senior Research Engineer


Yan is a Research Engineer in the Research Engineering and Design Lab (RED) at Adobe Research. He is primarily focused on designing, pruning, distilling, and conducting neural architecture search to develop efficient neural network structures and to productionize state-of-the-art deep learning techniques.

Prior to joining Adobe, Yan was a senior machine learning engineer at Samsung and received MSc in Computer Science from Rutgers University. He has hands-on industrial experience in optimizing and deploying a broad spectrum of computer vision tasks, including large-scale generative models (Diffusion), panoptic segmentation, semantic segmentation, image inpainting, low-light vision, etc. At Adobe, he is the tech lead of FireFly model efficiency to reduce model serving cost and to optimize Diffusion model architecture. He also collaborated closely with the Photoshop and Lightroom teams to streamline inpainting/masking related heavy-duty models and make them accessible across various platforms (PC, Mac, mobile device, in browser, etc.) where computation resources are limited.


SNED: Superposition Network Architecture Search for Efficient Video Diffusion Model

Li, Z., Kang, Y., Liu, Y., Liu, D., Hinz, T., Liu, F., Wang, Y. (Jun. 19, 2024)

CVPR 2024

AT-EDM: Attention-Driven Training-Free Efficiency Enhancement of Diffusion Models

Wang, H., Liu, D., Kang, Y., Li, Y., Lin, Z., Jha, N., Liu, Y. (Jun. 17, 2024)

CVPR 2024