Google logo

How Google Uses Reinforcement Learning

7 engineering articles about Reinforcement Learning from Google's engineering team

Articles

Filter:
Google logo
Google
Advanced
TorchTPU integrates a native torch.device("tpu") backend directly into PyTorch, allowing developers to train large-scale models on Google Cloud TPUs without rewriting code in JAX. The TorchTitan TPU fork enhances this with production-grade scaling primitives like SimpleFSDP, Graph Trainer, and custom Pallas kernels that achieve high Model Flops Utilization (MFU) across dense and MoE architectures. Developers can leverage these battle-tested optimization playbooks, recipes, and the Hugging Face modeling backend to efficiently scale their own workloads on TPU infrastructure.
Aleksey Vlasenko, Rob Torres, Will Constable
22 min read
Includes Code
--
Google logo
Google
Advanced
Imagine going to sleep after writing a single Markdown specification and waking up to find that an A...
Wei Wei
4 min read
Includes Code
--
Google logo
Google
Advanced
Tunix is Google’s new JAX-native post-training library designed to eliminate TPU idling bottlenecks when training multi-turn, tool-using LLM reasoning agents. It maximizes hardware throughput by combining highly concurrent, asynchronous rollouts with a decoupled producer-consumer pipeline, ensuring the trainer is constantly fed even while agents wait on network I/O or environment steps. Additionally, Tunix provides plug-and-play abstractions and continuous macro-level profiling, allowing developers to easily integrate custom open-source environments and optimize complex distributed workflows without massive code rewrites.
Haoyu Gao, Lance Wang, Shadi Noghabi, Tianshu Bao, Weiren Yu
10 min read
Includes Code
--
Google logo
Google
Advanced
The Google Tunix Hackathon on Kaggle challenged developers to transform small, non-reasoning base models into general reasoning engines using Kaggle TPUs and a limited compute budget. The winning teams achieved this by implementing multi-stage post-training pipelines that combined Supervised Fine-Tuning (SFT) with advanced alignment techniques like GRPO and SimPO. Ultimately, the competition democratized AI development by proving that highly capable, structured reasoning models can be successfully trained by the community using accessible, open-source resources.
Wei Wei, Weiren Yu, Tianshu Bao, Lance Wang, Chris Achard
6 min read
Includes Code
--
Google logo
Google
Intermediate
MaxText has introduced new support for Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) on single-host TPU configurations, leveraging JAX and the Tunix library for high-performance model refinement. These features enable developers to easily adapt pre-trained models for specialized tasks and complex reasoning using efficient algorithms like GRPO and GSPO. This update streamlines the post-training workflow, offering a scalable path from single-host setups to larger multi-host configurations.
Wei Wei, Weiren Yu
3 min read
Includes Code
--
Google logo
Google
Advanced
The article introduces Tunix, a new open-source, JAX-native library designed for post-training of large language models (LLMs).
Srikanth Kilaru, Tianshu Bao
7 min read
Includes Code
Has Summary
--
Google logo
Google
Intermediate
Gemma 3 is the latest version of the Gemma open-model family, boasting enhanced capabilities such as multimodality, longer context windows, and improved reasoning.
Omar Sanseviero, Philipp Schmid
5 min read
Includes Code
Has Summary
--

You've reached the end! All 7 articles loaded.