Sotabase

Career

· Principal Architect, NVIDIA2024–
· Adjunct Professor Of Computer Science, Stanford University2023–2023
· Director AI Architecture, Samsung Electronics America2020–2024
· Adjunct Professor Of Computer Science, Stanford University2020–2020
· Lead Member Of Technical Staff (Next Gen), Cerebras Systems2018–2020
· Adjunct Professor Of Computer Science, Stanford University2018–2020
· Member Of Technical Staff, Cerebras Systems2016–2018
· Principal Research Scientist, Movidius2015–2016
· Research Fellow, Stanford University2014–
· HPC Research Intern, Texas Advanced Computing Center (TACC)2013–2013
· Postdoctoral Research Fellow, The University of Texas at Austin2013–2014

Publications (24)

Plasticine: A reconfigurable architecture for parallel patterns
International Symposium on Computer Architecture · 2017
267
cited
Spatial: a language and compiler for application accelerators
ACM-SIGPLAN Symposium on Programming Language Design and Implementation · 2018
197
cited
Asynchrony begets momentum, with an application to deep learning
Allerton Conference on Communication, Control, and Computing · 2016
150
cited
Caffe con Troll: Shallow Ideas to Speed Up Deep Learning
DanaC@SIGMOD · 2015
73
cited
Omnivore: An Optimizer for Multi-device Deep Learning on CPUs and GPUs
arXiv.org · 2016
66
cited
Impact of FPGA architecture on resource sharing in high-level synthesis
Symposium on Field Programmable Gate Arrays · 2012
60
cited
Automating the Design of Processor/Accelerator Embedded Systems with LegUp High-Level Synthesis
2014 12th IEEE International Conference on Embedded and Ubiquitous Computing · 2014
29
cited
Plasticine: A Reconfigurable Accelerator for Parallel Patterns
IEEE Micro · 2018
18
cited
Profiling-driven multi-cycling in FPGA high-level synthesis
Design, Automation and Test in Europe · 2015
13
cited
Importance Sampling over Sets: A New Probabilistic Inference Scheme
Conference on Uncertainty in Artificial Intelligence · 2015
10
cited
TensorFlow to Cloud FPGAs: Tradeoffs for Accelerating Deep Neural Networks
International Conference on Field-Programmable Logic and Applications · 2019
9
cited
Techniques for mapping deep neural network frameworks to programmable accelerators
2021
Sotabase
Stefan Hadjis | Researcher Profile | Sotabase | Sotabase