Hao Zhuang, Ph.D., UCSD Computer Science
Currently, I solve problems ranging from software to hardware/silicon chip designs, and lead a team at Google, mainly working on media accelerators, customized compute, AI perception, and high-performance computing. My team at Google also maintains software package QKeras with AutoQKeras for ML software/hardware co-design, which is also supported by other open-sourced software toolchains to convert quantized ML models to hardware implementations such as FPGA and ASIC.
Previously, I applied math and computer science to model and analyze dynamic systems inside large-scale integrate circuits (VLSI chips). I also did a little bit of place-and-route in VLSI physical design. My research works and software packages have been transferred and used in the EDA industry.
Interests:
Algorithms, computer architecture, software/hardware co-design, compilers, EDA, design methodologies and automation for computing systems and accelerators. And their applications of
Computer vision, video compression, video understanding, and image processing.
Matrix and numerical computing, differential equations, sparse matrices, graphs, random walk mehtods.
Simulation and optimzation of dynamical systems, parallel computing, trading, and high-performance computing (HPC).
Services
Memberships
IEEE student memeber 2011-2016
ACM student memeber 2014-2016.
SIAM student memeber 2014-2016.
Teaching
|