Jeff Smith


I'm an AI R&D leader, focused on what's next for science and industry. 


I most recently worked at Meta Reality Labs on EMG-based neuromotor interfaces. I joined originally to partner with the creators of PyTorch and grow it from a research tool to the dominant deep learning framework in the field. Along the way, I helped shepherd Papers with Code to Meta/Facebook AI. 


As part of my research leadership at FAIR, I was privileged to support the researchers behind the following projects:

Open Source

I've been involved with a range of different open source projects, primarily focused on tools for AI/ML. During my team leading the PyTorch core development team at Facebook's FAIR (now Meta), I was mostly deeply involved in the development of our domain libraries for Vision, Text, and Audio, as well as our work to scale training via distribution. I'm particularly proud of bringing my production experience to the creation of TorchServe with Amazon and the community. Finally, the collaborations that led to the CrypTen and VISSL libraries were excellent examples of how PyTorch still powers research.

Beyond PyTorch, I've been involved with the Elixir ML open source ecosystem, Axon, Nx, and my own work around ONNX and Elixir.

Prior to that, most of my open source work was focused on the JVM ML ecosystem, primarily focused on Spark. This work was mostly captured in my first book.

Writing and Speaking

I've written and spoken about a lot of this work, and this sort organizes some of that content. It contains some links to my work across writing software and writing about software


I've founded, led, and helped many startups. Check LinkedIn, for more detail.


I'm always interested in what's next in AI, so please reach out, if you'd like to connect on any of these topics.