Igor Fedorov

Pod / Tech Lead & AI Research Scientist, Meta

I am an AI Research Scientist at Meta, where I am a tech lead for large language model compression and inference optimization for edge use cases. I have also been responsible for algorithm research and software development for neural architecture search on ads recommendation models. Before that, I was a Staff Research Engineer in the machine learning research group at ARM Research. I completed my PhD in Electrical Engineering in 2018 at the University of California San Diego, supervised by Prof. Bhaskar Rao and co-advised by Prof. Truong Nguyen. Before UCSD, I completed my B.S. and M.S. (supervised by Prof. Pierre Moulin) in Electrical Engineering at the University of Illinois Urbana-Champaign.

I am lucky to be married to Katarina Fedorov.

Igor Fedorov

Research & Patents

40 entries

MobileLLM-Flash: Latency-Guided On-Device LLM Design for Industry Scale Deployment

Hanxian Huang*, Igor Fedorov*, Andrey Gromov, Bernard Beckerman, Naveen Suda, David Eriksson, Maximilian Balandat, Rylan Conway, Patrick Huber, Chinnadhurai Sankar, Ayushi Dalmia, Zechun Liu, Lemeng Wu, Tarek Elgamal, Adithya Sagar, Vikas Chandra, Raghuraman Krishnamoorthi

ACL 2026

Short Data, Long Context: Distilling Positional Knowledge in Transformers

Patrick Huber, Ernie Chang, Chinnadhurai Sankar, Rylan Conway, Igor Fedorov, Md Rifat Arefin, Adithya Sagar

arXiv 2026

MobileLLM-Pro Technical Report

Patrick Huber, Ernie Chang, Wei Wen, Igor Fedorov, Tarek Elgamal, Hanxian Huang, Naveen Suda, Chinnadhurai Sankar, Vish Vogeti, Yanghan Wang, Alex Gladkov, Kai Sheng Tai, Abdelrahman Elogeel, Tarek Hefny, Vikas Chandra, Ahmed Aly, Anuj Kumar, Raghuraman Krishnamoorthi, Adithya Sagar

arXiv 2025
SpinQuant

SpinQuant: LLM Quantization with Learned Rotations

Zechun Liu, Changsheng Zhao, Igor Fedorov, Bilge Soran, Dhruv Choudhary, Raghuraman Krishnamoorthi, Vikas Chandra, Yuandong Tian, Tijmen Blankevoort

ICLR 2025
Llama Guard 3-1B-INT4

Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations

Igor Fedorov, Kate Plawiak, Lemeng Wu, Tarek Elgamal, Naveen Suda, Eric Smith, Hongyuan Zhan, Jianfeng Chi, Yuriy Hulovatyy, Kimish Patel, Zechun Liu, Changsheng Zhao, Yangyang Shi, Tijmen Blankevoort, Mahesh Pasupuleti, Bilge Soran, Zacharie Delpierre Coudert, Rachad Alao, Raghuraman Krishnamoorthi, Vikas Chandra

arXiv 2024
MobileLLM

MobileLLM: Optimizing Sub-billion Parameter Language Models for On-Device Use Cases

Z. Liu, C. Zhao, F. Iandola, C. Lai, Y. Tian, I. Fedorov, Y. Xiong, E. Chang, Y. Shi, R. Krishnamoorthi, L. Lai, V. Chandra

ICML 2024
Dictionaries in machine learning

Dictionaries in Machine Learning

K. Kreutz-Delgado, B. D. Rao, I. Fedorov, S. Das

Signal Processing and Machine Learning Theory, 2024
Rankitect

Rankitect: Ranking Architecture Search Battling World-class Engineers at Meta Scale

W. Wen, K. Liu, I. Fedorov, X. Zhang, H. Yin, W. Chu, K. Hassani, M. Sun, J. Liu, X. Wang, L. Jiang, Y. Chen, B. Zhang, X. Liu, D. Cheng, Z. Chen, G. Zhao, F. Han, J. Yang, Y. Hao, L. Xiong, W. Chen

arXiv 2023
SiGeo

SiGeo: Sub-One-Shot NAS via Information Theory and Geometry of Loss Landscape

H. Zheng, K. Liu, I. Fedorov, X. Zhang, W. Chen, W. Wen

arXiv 2023
DistDNAS

DistDNAS: Search Efficient Feature Interactions within 2 Hours

T. Zhang, W. Wen, I. Fedorov, X. Liu, B. Zhang, F. Han, W. Chen, Y. Han, F. Yan, H. Li, Y. Chen

arXiv 2023
PerfSAGE

PerfSAGE: Generalized Inference Performance Predictor for Arbitrary Deep Learning Models on Edge Devices

Y. Chai, D. Tripathy, C. Zhou, D. Gope, I. Fedorov, R. Matas, D. Brooks, G. Wei, P. Whatmough

arXiv 2023
Efficient Edge Inference by Selective Query

Efficient Edge Inference by Selective Query

A. Kag, I. Fedorov, A. Gangrade, P. N. Whatmough, V. Saligrama

ICLR 2023

Error Detection

M. Haddon, I. Fedorov, R. Jeyapaul, P. N. Whatmough, Z. Liu

US Patent Application, 2022
UDC

UDC: Unified DNAS for Compressible TinyML Models

I. Fedorov, R. M. Navarro, H. Tann, C. Zhou, M. Mattina, P. N. Whatmough

NeurIPS 2022
Restructurable Activation Networks

Restructurable Activation Networks

K. Bhardwaj, J. Ward, C. Tung, D. Gope, L. Meng, I. Fedorov, A. Chalfin, P. Whatmough, D. Loh

arXiv 2022
Selective Cloud Interactions

Achieving High TinyML Accuracy through Selective Cloud Interactions

A. Kag, I. Fedorov, A. Gangrade, P. N. Whatmough, V. Saligrama

ICML DyNN Workshop, 2022

Neural Network System and Training Method

I. Fedorov, P. Whatmough

US Patent Application, 2022

A Unified Neural Network Optimization Framework

I. Fedorov, R. Matas, C. Zhou, H. Tann, P. Whatmough, M. Mattina

US Patent Application, 2022

System, Devices and/or Processes for Defining a Search Space for Neural Network Processing Device Architectures

H. Tann, R. Navarro, I. Fedorov, C. Zhou, P. Whatmough, M. Mattina

US Patent Application
MicroNets

MicroNets: Neural Network Architectures for Deploying TinyML Applications on Commodity Microcontrollers

C. Banbury*, C. Zhou*, I. Fedorov*, R. M. Navarro, U. Thakker, D. Gope, V. J. Reddi, M. Mattina, P. N. Whatmough

MLSys 2021

Image Denoising Neural Network Architecture and Method of Training the Same

M. El-Khamy, I. Fedorov, J. Lee

US Patent 2020
TinyLSTMs

TinyLSTMs: Efficient Neural Speech Enhancement for Hearing Aids

I. Fedorov, M. Stamenovic, C. Jensen, L. Yang, A. Mandell, Y. Gan, M. Mattina, P. N. Whatmough

INTERSPEECH 2020
SSGD

SSGD: Sparsity-promoting Stochastic Gradient Descent Algorithm for Unbiased DNN Pruning

C. Lee, I. Fedorov, B. D. Rao, H. Garudadri

ICASSP 2020
Mango

Mango: A Python Library for Parallel Hyperparameter Tuning

S. Sandha, M. Aggarwal, I. Fedorov, M. Srivastava

ICASSP 2020
SpArSe

SpArSe: Sparse Architecture Search for CNNs on Resource-Constrained Microcontrollers

I. Fedorov, R. P. Adams, M. Mattina, P. N. Whatmough

NeurIPS 2019

Pushing the Limits of RNN Compression

U. Thakker, I. Fedorov, J. Beu, D. Gope, C. Zhou, G. Dasika, M. Mattina

EMC2 Workshop, 2019

Compressing RNNs for IoT Devices by 15-38x using Kronecker Products

U. Thakker, J. Beu, D. Gope, C. Zhou, I. Fedorov, G. Dasika, M. Mattina

arXiv 2019

Structured Learning with Scale Mixture Priors

I. Fedorov

PhD Thesis
Multimodal Sparse Bayesian Dictionary Learning

Multimodal Sparse Bayesian Dictionary Learning

I. Fedorov, B. D. Rao

arXiv preprint
MSBDL for Multimodal Data Classification

Multimodal Sparse Bayesian Dictionary Learning Applied to Multimodal Data Classification

I. Fedorov, B. D. Rao, T. Q. Nguyen

ICASSP 2016
Re-weighted Learning for Sparsifying DNNs

Re-weighted Learning for Sparsifying Deep Neural Networks

I. Fedorov, B. D. Rao

arXiv preprint, 2018
Sparse Non-Negative Least Squares

A Unified Framework for Sparse Non-Negative Least Squares using Multiplicative Updates and the Non-Negative Matrix Factorization Problem

I. Fedorov, A. Nalci, R. Giri, B. D. Rao, T. Q. Nguyen, H. Garudadri

Signal Processing, 2018
Rectified Gaussian Scale Mixtures

Rectified Gaussian Scale Mixtures and the Sparse Non-Negative Least Squares Problem

A. Nalci, I. Fedorov, M. Al-Shoukairi, T. T. Liu, B. D. Rao

IEEE Transactions on Signal Processing, 2018
Relevance Subject Machine

Relevance Subject Machine: A Novel Person Re-Identification Framework

I. Fedorov, R. Giri, B. D. Rao, T. Q. Nguyen

arXiv preprint, 2017
Robust Bayesian Block Sparse Signal Recovery

Robust Bayesian Method for Simultaneous Block Sparse Signal Recovery with Applications to Face Recognition

I. Fedorov, R. Giri, B. D. Rao, T. Q. Nguyen

ICIP 2016
Image Reconstruction under Imaging Time Constraints

Image Reconstruction under Imaging Time Constraints

I. Fedorov, S. Obrzut, B. Song, B. D. Rao

Asilomar Conference, 2017

Total Variation Regularization in I-123 Ioflupane SPECT Reconstruction

I. Fedorov, B. Song, B. D. Rao, I. Levitan, S. Obrzut

Journal of Nuclear Medicine, 2017
Kinect depth video compression

Kinect Depth Video Compression for Action Recognition

I. Fedorov

M.S. Thesis, 2014

Automated Worker Activity Analysis in Indoor Environments for Direct-Work Rate Improvement from Long Sequences of RGB-D Images

A. Khosrowpour, I. Fedorov, A. Holynski, J. C. Niebles, M. Golparvar-Fard

Construction Research Congress, 2014

Power Delivery for Series Connected Voltage Domains in Digital Circuits

P. Shenoy, I. Fedorov, T. Neyens, P. Krein

Intl. Conference on Energy Aware Computing, 2011