Posts by Tianshu Bao

4 results

Clear filters
  • NOV. 16, 2026 / AI

    Unlocking Stable On-Policy RL on TPU with Bitwise-Zero Train–Inference Mismatch

    How we eliminated floating-point train–inference logprob drift for on-policy LLM reinforcement learning across 64 Cloud TPU v5p chips using a generalizable four-part JAX and Pallas recipe in Tunix.

    zero_tim_tpu_blog_header_1600x900
  • JULY 21, 2026 / AI

    Scaling Agentic RL: High-Throughput Agentic Training with Tunix

    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.

    Banner Image
  • MAY 28, 2026 / AI

    How the community trained Gemma to "Think" with Tunix and TPUs

    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.

    Building-1-banner
  • SEPT. 30, 2025 / AI

    Introducing Tunix: A JAX-Native Library for LLM Post-Training

    Tunix is a new JAX-native, open-source library for LLM post-training. It offers comprehensive tools for aligning models at scale, including SFT, preference tuning (DPO), advanced RL methods (PPO, GRPO, GSPO), and knowledge distillation. Designed for TPUs and seamless JAX integration, Tunix emphasizes developer control and shows a 12% relative improvement in pass@1 accuracy on GSM8K.

    Tunix logo