Sotabase

Career

· Assistant Professor, New York University2025–
· Assistant Professor, New York University (NYU) Tandon School of Engineering2025–
· Research Scientist, Facebook AI Infrastructure Research
· PhD in Computer Science, Harvard University
· Undergraduate Degrees in Computer Systems Engineering and Applied Mathematics, University of Massachusetts, Amherst

Publications (67)

Minerva: Enabling Low-Power, Highly-Accurate Deep Neural Network Accelerators
International Symposium on Computer Architecture · 2016
602
cited
Machine Learning at Facebook: Understanding Inference at the Edge
International Symposium on High-Performance Computer Architecture · 2019
467
cited
Ares: A framework for quantifying the resilience of deep neural networks
Design Automation Conference · 2018
329
cited
MachSuite: Benchmarks for accelerator design and customized architectures
IEEE International Symposium on Workload Characterization · 2014
317
cited
The Architectural Implications of Facebook's DNN-Based Personalized Recommendation
International Symposium on High-Performance Computer Architecture · 2019
316
cited
RecNMP: Accelerating Personalized Recommendation with Near-Memory Processing
International Symposium on Computer Architecture · 2019
251
cited
DeepRecSys: A System for Optimizing End-To-End At-Scale Neural Recommendation Inference
International Symposium on Computer Architecture · 2020
200
cited
Fathom: reference workloads for modern deep learning methods
IEEE International Symposium on Workload Characterization · 2016
186
cited
RETROSPECTIVE: Aladdin: a Pre-RTL, Power-Performance Accelerator Simulator Enabling Large Design Space Exploration of Customized Architectures
2023
180
cited
Cheetah: Optimizing and Accelerating Homomorphic Encryption for Private Inference
International Symposium on High-Performance Computer Architecture · 2020
172
cited
DeepReDuce: ReLU Reduction for Fast Private Inference
International Conference on Machine Learning · 2021
103
cited
CryptoNAS: Private Inference on a ReLU Budget
Neural Information Processing Systems · 2020
99
cited
A case for efficient accelerator design space exploration via Bayesian optimization
International Symposium on Low Power Electronics and Design · 2017
83
cited
Selective Network Linearization for Efficient Private Inference
International Conference on Machine Learning · 2022
51
cited
Porcupine: a synthesizing compiler for vectorized homomorphic encryption
ACM-SIGPLAN Symposium on Programming Language Design and Implementation · 2021
50
cited
MASR: A Modular Accelerator for Sparse RNNs
International Conference on Parallel Architectures and Compilation Techniques · 2019
48
cited
The Aladdin Approach to Accelerator Design and Modeling
IEEE Micro · 2015
43
cited
Weightless: Lossy Weight Encoding For Deep Neural Network Compression
International Conference on Machine Learning · 2017
41
cited
Circa: Stochastic ReLUs for Private Deep Learning
Neural Information Processing Systems · 2021
38
cited
Orion: A Fully Homomorphic Encryption Framework for Deep Learning
International Conference on Architectural Support for Programming Languages and Operating Systems · 2023
34
cited
Sotabase