Jeff Johnson
Meta FAIR research scientist; implemented the GPU version of the Faiss similarity search library
Jeff Johnson is a research scientist in the SysML research group at Meta's Fundamental AI Research (FAIR) lab, working on GPU-based algorithms for machine learning and on ASIC and FPGA computer arithmetic for linear algebra. The Faiss README credits Johnson with implementing all of the GPU version of Faiss, and Johnson is first author of "Billion-scale similarity search with GPUs" (2017). Johnson's GitHub profile also says Johnson wrote the original PyTorch GPU backend.1342
Research
Johnson works in FAIR's SysML research group on GPU algorithms for machine learning and on ASIC and FPGA arithmetic for linear algebra. Johnson's papers include "Rethinking floating point for deep learning" (December 2018) and "Efficient, arbitrarily high precision hardware logarithmic arithmetic for linear algebra" (May 2020), both as sole author.1
Faiss
Work
FaissOpen-source code
Credited in the Faiss README with implementing all of the GPU Faiss, and first author of the 2017 paper that the README asks users to cite for GPU Faiss.34
PyTorch GPU backend (external site: github.com)Open-source code
Johnson's GitHub profile states that Johnson wrote the original PyTorch GPU backend.2
Career timeline
Feb 2017
Posted "Billion-scale similarity search with GPUs" as first author, alongside the Faiss release.43
Dec 2018
Published "Rethinking floating point for deep learning" as sole author.1
May 2020
Published work on high-precision hardware logarithmic arithmetic for linear algebra.1
Jan 2024
Co-authored "The Faiss library" with Matthijs Douze and others.5
Related people
More sourced profiles at the same organization, in the same field, or from the same school.
- Adina WilliamsAlso at Meta
- Gavin UbertiChips and hardware
- Alban DesmaisonAlso at Meta
- Ian BuckChips and hardware
- Alexandr WangAlso at Meta
- Johannes GäßlerInference and serving
- Alexandre GramfortAlso at Meta
- Aaron TeoInference and serving
- Andrew BosworthAlso at Meta
- Andrew FeldmanChips and hardware
- Christoph FeichtenhoferAlso at Meta Superintelligence Labs
- Awni HannunML frameworks and libraries
Is this you?
Profiles cover public professional work only, and every line cites a source. If something is out of date, missing or wrong, send a correction with a link to an official page and we will update it. Suggest a correction
Add this badge to your website or GitHub profile README. It links back to this page. It is not an endorsement or a rating.
Markdown (for a README)
[](https://unitedstatesofamericasuperintelligence.com/people/jeff-johnson/)
HTML
<a href="https://unitedstatesofamericasuperintelligence.com/people/jeff-johnson/"><img src="https://unitedstatesofamericasuperintelligence.com/badges/profile.svg" alt="Jeff Johnson — Profile on USASI" height="20"></a>
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.