Jongse Park

Georgia Institute of Technology, Kootenay Association for Science & Technology

Papers

Total Citations

223

H-Index

About

Jongse Park is a computer architecture and systems researcher whose work sits at the intersection of hardware acceleration, energy efficiency, and machine learning systems. He is best known for his pioneering contributions to analog computing as a means of accelerating general-purpose code, a radical departure from conventional digital processing approaches. His landmark work, "General-purpose code acceleration with limited-precision analog computation" (2014), proposed a holistic solution spanning circuits to compilers that enables energy-efficient computation through limited-precision analog techniques — a contribution that has garnered over 200 cumulative citations and remains highly influential in the approximate computing community. This research addressed the growing challenge of diminishing returns in per-transistor performance improvements, offering a compelling architectural alternative at a time when the field urgently needed new directions. More recently, Park has extended his expertise toward the demands of autonomous systems, with his 2024 work on DACAPO tackling the formidable challenge of continuous deep neural network learning for real-time video analytics on resource-constrained platforms such as self-driving vehicles and drones. His career reflects a consistent drive to bridge theoretical innovation with practical, systems-level impact across evolving computing paradigms.