PeopleofSuperintelligence

Builder

Natalia Gimelshein

Natalia Gimelshein is a PyTorch maintainer and engineer whose work connects GPU memory management, framework performance, and compilation. She coauthored vDNN in 2016 and the PyTorch library paper in 2019, then coauthored the 2021 announcement of native CUDA Graph execution. At a January 2023 Q&A, she spoke about TorchInductor and PT2 backend integration; in 2024, she coauthored the PyTorch 2 paper and received a PyTorch Problem Solver award. She worked as a software engineer at Meta in 2025 and, as of October 2026, is a PyTorch Module Maintainer and a software engineer at Meta.

Works
73
Citations
33,578
h-index
18
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2016

    vDNN and GPU memory

    In 2016, she coauthors vDNN, a runtime memory manager that uses CPU and GPU memory together to train larger neural networks.

  2. 2019

    The PyTorch library paper

    In 2019, she coauthors the NeurIPS paper that explains PyTorch's architecture, imperative Python style, and efficient support for GPUs.

  3. 2021

    CUDA Graphs enter PyTorch

    In 2021, she works on GPU performance optimization for deep-learning workloads at NVIDIA and Facebook as an Applied Research Scientist at Facebook. She is a PyTorch core-team member and coauthors the announcement that brings native CUDA Graph execution to PyTorch.

  4. January 2023

    TorchInductor backend Q&A

    At the January 2023 PyTorch 2.0 Q&A, she speaks in a session on TorchInductor and PT2 backend integration.

  5. 2024

    The PyTorch 2 paper

    In 2024, she coauthors the PyTorch 2 paper on dynamic Python bytecode transformation and graph compilation, with OpenAI as her affiliation.

  6. 2024

    PyTorch Problem Solver award

    In 2024, the PyTorch Foundation names her a Problem Solver in its contributor awards.

  7. 2025

    Software engineer at Meta

    In 2025, she works as a software engineer at Meta.

  8. As of October 2026

    Software engineer at Meta

    As of October 2026, she is a software engineer at Meta and a PyTorch Module Maintainer.

Career

  1. 2024OpenAI
  2. 2021NVIDIA
  3. 2021MetaApplied research scientist
  4. 2007–2013ERC Inc.
  5. 2007–2014United States Department of the Air Force
  6. 2001–2004Pennsylvania State University

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. PyTorch: An Imperative Style, High-Performance Deep Learning Library2019 · 31,488 citations
  2. PyTorch 2: Faster Machine Learning Through Dynamic Python Bytecode Transformation and Graph Compilation2024 · ACM International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) · 728 citationsWith Horace He
  3. vDNN: Virtualized deep neural networks for scalable, memory-efficient neural network design2016 · International Symposium on Microarchitecture · 416 citations
  4. Vibrational relaxation rates in the direct simulation Monte Carlo method2002 · Physics of Fluids · 114 citations
  5. Radiometric phenomena: From the 19th to the 21st century2012 · Vacuum · 95 citations

From OpenAlex, as of 21 October 2026.

  1. 22 of 133 people shaping AI with a former OpenAI role now work at AnthropicIncludes Natalia Gimelshein. Our estimates of where OpenAI recruits came from and where 133 people with former OpenAI roles work now, including the companies and labs they founded.
  2. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Natalia Gimelshein. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.
  3. Among employers with 30+ full-time cases, OpenAI leads primary technical role H-1B pay at $310,000, and Amazon, Google and Microsoft file 57% of all casesIncludes OpenAI. Across 440,669 certified US H-1B filings from 44 employers (FY2020–FY2026), OpenAI leads primary technical role base pay at $310,000 and xAI leads overall base pay at $290,000.
  4. Google has the highest combined h-index among AI organizationsIncludes NVIDIA. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Natalia Gimelshein yet. Send us a free-licensed photo.

  1. vDNN: Virtualized Deep Neural Networks for Scalable, Memory-Efficient Neural Network Design arxiv.org. Checked 21 October 2026.
  2. PyTorch: An Imperative Style, High-Performance Deep Learning Library proceedings.neurips.cc. Checked 21 October 2026.
  3. Accelerating PyTorch with CUDA Graphs pytorch.org. Checked 21 October 2026.
  4. PyTorch 2.0 Ask the Engineers Live Q&A Series: A deep dive on TorchInductor and PT2 Backend Integration pytorch.org. Checked 21 October 2026.
  5. Main Program – ASPLOS 2024 asplos-conference.org. Checked 21 October 2026.
  6. Announcing the 2024 PyTorch Contributor Awards pytorch.org. Checked 21 October 2026.
  7. Meet the Committees | PyTorch Conference 2025 events.linuxfoundation.org. Checked 21 October 2026.
  8. PyTorch Conference Europe 2026: A Landmark Moment for Open Source AI in Paris pytorch.org. Checked 21 October 2026.
  9. NVIDIA at PyTorch Conference 2026 nvidia.com. Checked 21 October 2026.