Soroush Ghodrati
Personal Webpage
About Me
My name is Soroush Ghodrati. I am a PhD student in Computer Science and Engineering (CSE) department under supervision of Prof. Hadi Esmaeilzadeh at UC San Diego.
My research interests lie at the intersection of Artificial Intelligence (AI), Computer Architecture, and System Design. Specifically, I am interested in devising new technologies and domain-specific computing architectures and systems to accelerate AI workloads execution.
My research is generously supported by a Google PhD Fellowship.
Research Interests:
Domain-Specific Computer Architectures and Systems for AI/ML
Computer Architecture
Deep Learning
VLSI
Education:
PhD in Computer Science and Engineering, in progress
Advisor: Prof. Hadi Esmaeilzadeh
UC San Diego
MS in Computer Science and Engineering, 2020
Advisor: Prof. Hadi Esmaeilzadeh
UC San Diego
BS in Electrical Engineering, 2017
Sharif University of Technology, Iran
Work Experiences:
Research Intern:
June 2022-Spet. 2022. Brain Team, Google Research
R&D Intern:
June 2021-Spet. 2021. Corporate R&D, Qualcomm
R&D Intern:
June 2020-Spet. 2020. Corporate R&D, Qualcomm
News:
June 2022: Excited to start my summer internship at Google Research Brain team!
March 2022: Our paper "Accelerating Attention through Gradient-Based Learned Runtime Pruning" has been accepted in International Symposium on Computer Architecture (ISCA) 2022.
October 2021: The Planaria's hardware simulator and task scheduler (our work published in MICRO'20) is now open sourced!
September 2021: Thrilled and honored to receive the 2021 Google PhD Fellowship in Systems and Networking!
October 2020: Our paper "A Computational Stack for Cross-Domain Acceleration" has been accepted in International Symposium on High Performance Computer Architecture (HPCA) 2021!
July 2020: Our paper "Mixed-Signal Charge-Domain Acceleration of Deep Neural Networks through Interleaved Bit-Partitioned Arithmetic" has been accepted in International Conference on Parallel Architectures and Compilation Techniques (PACT) 2020!
July 2020: Our paper "Planaria: Dynamic Architecture Fission for Spatial Multi-Tenant Acceleration of Deep Neural Networks" has been accepted in International Symposium on Microarchitecture (MICRO) 2020!
June 2020: Excited to start my summer internship as "ML Accelerator R&D intern" at Qualcomm, Corporate R&D division!
March 2020: Our team has been selected as one of the finalists for Qualcomm Innovation Fellowship 2020!
February 2020: Our paper "Bit-Parallel Vector Composability for Neural Acceleration" has been accepted in International Design Automation Conference (DAC) 2020!