Jean Kossaifi

Pioneering tensor methods and neural operators for scientific machine learning

Jean Kossaifi

About Me

I lead research at NVIDIA in the field of AI for Engineering and Scientific Simulation, where my work focuses on advancing new algorithmic paradigms to solve complex physics-based problems. My core research combines tensor methods with deep learning to develop efficient and powerful neural architectures.

A central part of my mission is to democratize advanced computational techniques. To that end, I created and lead the development to two widely-used open-source libraries: TensorLy, for tensor methods, and NeuralOperator, for scientific machine learning, helping to accelerate scientific discovery for the broader research community.

Prior to NVIDIA, I was a founding member of the Samsung AI Center in Cambridge. My academic foundation includes a French Engineering Diploma in Mathematics, Computer Science, and Finance and a BSc in advanced mathematics. I then completed my PhD in Artificial Intelligence at Imperial College London.

Research Interests

AI for Engineering Tensor Methods Neural Operators Scientific Machine Learning Deep Learning

Featured Work

A selection of projects and research I'm particularly proud of.

TensorLy: Democratizing Tensor Methods

TensorLy: Democratizing Tensor Methods

Open-Source Software

Created the leading Python library for tensor methods, now with 1,000+ GitHub stars and used by researchers worldwide. TensorLy provides a high-level API for tensor decomposition and tensorized neural networks.

ZenCFG

ZenCFG

Open-Source Software

Easy configuration of Python projects with minimal boilerplate. A clean, Pythonic approach to configuration management, particularly adapted for deep learning research.

Neural Operator

Neural Operator

Open-Source Software

Learning operators that map between function spaces for solving PDEs and scientific computing problems. A new paradigm for AI in science.

TensorLy-Torch

TensorLy-Torch

Deep tensorized neural networks in PyTorch. Provides out-of-the-box tensor layers, decomposition methods, and utilities for building efficient deep learning models.

Academic Service

Leadership & Editorial

2024 – Present

Peer Review

Area Chair

NeurIPS 2022 – 2024 ICML 2023 – 2025 ICLR 2023

Reviewer

TPAMI 2020 – 2021 JMLR 2017 - 2022 ICCV 2019 - 2021 SciPy 2021 NeurIPS 2018, 2020, 2021 ICLR 2019 - 2020 ICML 2019 CVPR Outstanding Reviewer 2020 2019 -2020 AAAI 2020 ECCV 2020 Transactions on Signal Processing 2021 - 2022 Image and Vision Computing Journal 2014 - 2018 IEEE Transactions on Emerging Topics in Computing 2019 IEEE Sensors 2022

Selected Publications

TensorGRaD: Tensor Gradient Robust Decomposition for Memory-Efficient Neural Operator Training

Sebastian Loeschcke, David Pitt, Robert Joseph George, Jiawei Zhao, Cheng Luo, Yuandong Tian, Jean Kossaifi, Anima Anandkumar

arXiv preprint arXiv:2501.02379, 2025

Neural operators for accelerating scientific simulations and design

Kamyar Azizzadenesheli, Nikola Kovachki, Zongyi Li, Miguel Liu-Schiaffini, Jean Kossaifi, Anima Anandkumar

Nature Reviews Physics, 2024

A library for learning neural operators

Jean Kossaifi, Nikola Kovachki, Zongyi Li, David Pitt, Miguel Liu-Schiaffini, Robert Joseph George, Boris Bonev, Kamyar Azizzadenesheli, Julius Berner, Valentin Duruisseaux, others

arXiv preprint arXiv:2412.10354, 2024

Estimation of continuous valence and arousal levels from faces in naturalistic conditions

Antoine Toisoul, Jean Kossaifi, Adrian Bulat, Georgios Tzimiropoulos, Maja Pantic

Nature Machine Intelligence, 2021

Factorized Higher-Order CNNs with an Application to Spatio-Temporal Emotion Estimation

Jean Kossaifi, Antoine Toisoul, Adrian Bulat, Yannis Panagakis, Maja Pantic

IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020

TensorLy: Tensor Learning in Python

Jean Kossaifi, Yannis Panagakis, Anima Anandkumar, Maja Pantic

Journal of Machine Learning Research (JMLR), 2019
TensorLy: Tensor Learning in Python

Spectral learning on matrices and tensors

Majid Janzamin, Rong Ge, Jean Kossaifi, Anima Anandkumar

Foundations and Trends{\textregistered} in Machine Learning, 2019

Recent News

May 2025

Guest speaker at the Oak Ridge National Laboratory's AI Expo, where I presented my work on Neural Operators for AI in Science and Engineering

2024

Co-authored a book chapter, "Tensor methods in deep learning", in the book Signal Processing and Machine Learning Theory

November 2024

Co-organized workshop with Topal team at INRIA Bordeaux on efficient scaling of neural architectures, covering re-materialization, offloading, scheduling and model pipelining

July 2024

Co-organizer, ICML Workshop on Advancing Neural Network Training (WANT): Computational Efficiency, Scalability, and Resource Optimization