Media Summary: Matthew Johnson, Principal Scientist at Google, delivers the Keynote address at the Relive your favorite moments from the Fall 2025 Matthew Watson, Fabien Hertschuh, and Abheesht Sharma explain how Keras can be used to run

Jax Openxla Devlab Fall 2025 - Detailed Analysis & Overview

Matthew Johnson, Principal Scientist at Google, delivers the Keynote address at the Relive your favorite moments from the Fall 2025 Matthew Watson, Fabien Hertschuh, and Abheesht Sharma explain how Keras can be used to run Jen Ha and Tom Yeh use spreadsheets to demonstrate the sharding in Qwen3. Distinguished Engineer, Robert Hundt, from Google gives the keynote speech for the Lin Chai, Lance Wang, and Abheesht Sharma from Google talk about the latest features in the framework for post-training in

David Hall gives the latest updates on Marin. Minho Ryu, a Google Developer Expert in AI, talks about how you can contribute to the Jiaxin Cao shows off how to execute async PjRT on GPU. Kyle Meggs talks about the latest version of MaxText Sharad Vikram explains how to use Pallas to make your custom kernels on TPUs go fast!

Photo Gallery

Keynote address with Matt Johnson | JAX/OpenXLA DevLab Fall 2025
JAX on GPUs | JAX/OpenXLA DevLab Fall 2025
JAX/OpenXLA DevLab Fall 2025 Event Recap
Keras: Deep Learning Framework for JAX | JAX/OpenXLA DevLab Fall 2025
AMD's JAX Ecosystem | JAX/OpenXLA DevLab Fall 2025
Models in JAX & NNX by Hand | JAX/OpenXLA DevLab Fall 2025
JAX/OpenXLA DevLab 2025 Keynote Speech with Robert Hundt
Tunix: Tune in JAX | JAX/OpenXLA DevLab Fall 2025
Marin: An Open Laboratory for Foundation Models in JAX | JAX/OpenXLA DevLab Fall 2025
Contributing to the JAX Ecosystem | JAX/OpenXLA DevLab Fall 2025
Asynchronous PjRT for JAX on GPU | JAX/OpenXLA DevLab Fall 2025
RL at Google Scale, with MaxText 2.0 | JAX/OpenXLA DevLab Fall 2025
View Detailed Profile
Keynote address with Matt Johnson | JAX/OpenXLA DevLab Fall 2025

Keynote address with Matt Johnson | JAX/OpenXLA DevLab Fall 2025

Matthew Johnson, Principal Scientist at Google, delivers the Keynote address at the

JAX on GPUs | JAX/OpenXLA DevLab Fall 2025

JAX on GPUs | JAX/OpenXLA DevLab Fall 2025

Kuy Mainwaring shows how to run

JAX/OpenXLA DevLab Fall 2025 Event Recap

JAX/OpenXLA DevLab Fall 2025 Event Recap

Relive your favorite moments from the Fall 2025

Keras: Deep Learning Framework for JAX | JAX/OpenXLA DevLab Fall 2025

Keras: Deep Learning Framework for JAX | JAX/OpenXLA DevLab Fall 2025

Matthew Watson, Fabien Hertschuh, and Abheesht Sharma explain how Keras can be used to run

AMD's JAX Ecosystem | JAX/OpenXLA DevLab Fall 2025

AMD's JAX Ecosystem | JAX/OpenXLA DevLab Fall 2025

Andy Ye talks about how AMD is using

Models in JAX & NNX by Hand | JAX/OpenXLA DevLab Fall 2025

Models in JAX & NNX by Hand | JAX/OpenXLA DevLab Fall 2025

Jen Ha and Tom Yeh use spreadsheets to demonstrate the sharding in Qwen3.

JAX/OpenXLA DevLab 2025 Keynote Speech with Robert Hundt

JAX/OpenXLA DevLab 2025 Keynote Speech with Robert Hundt

Distinguished Engineer, Robert Hundt, from Google gives the keynote speech for the

Tunix: Tune in JAX | JAX/OpenXLA DevLab Fall 2025

Tunix: Tune in JAX | JAX/OpenXLA DevLab Fall 2025

Lin Chai, Lance Wang, and Abheesht Sharma from Google talk about the latest features in the framework for post-training in

Marin: An Open Laboratory for Foundation Models in JAX | JAX/OpenXLA DevLab Fall 2025

Marin: An Open Laboratory for Foundation Models in JAX | JAX/OpenXLA DevLab Fall 2025

David Hall gives the latest updates on Marin.

Contributing to the JAX Ecosystem | JAX/OpenXLA DevLab Fall 2025

Contributing to the JAX Ecosystem | JAX/OpenXLA DevLab Fall 2025

Minho Ryu, a Google Developer Expert in AI, talks about how you can contribute to the

Asynchronous PjRT for JAX on GPU | JAX/OpenXLA DevLab Fall 2025

Asynchronous PjRT for JAX on GPU | JAX/OpenXLA DevLab Fall 2025

Jiaxin Cao shows off how to execute async PjRT on GPU.

RL at Google Scale, with MaxText 2.0 | JAX/OpenXLA DevLab Fall 2025

RL at Google Scale, with MaxText 2.0 | JAX/OpenXLA DevLab Fall 2025

Kyle Meggs talks about the latest version of MaxText

Pallas TPU: New and Advanced Features for Kernels | JAX/OpenXLA DevLab Fall 2025

Pallas TPU: New and Advanced Features for Kernels | JAX/OpenXLA DevLab Fall 2025

Sharad Vikram explains how to use Pallas to make your custom kernels on TPUs go fast!