Open-source AI inference and training infrastructure
radixark.com (opens in a new tab)RadixArk builds open systems for AI inference and training, with the goal of making frontier-grade AI infrastructure widely available. Its engineers created SGLang, a popular open-source engine for serving large language models with more than 30,000 GitHub stars, and developed Miles, a framework for large-scale reinforcement learning. Its founders previously worked on AI infrastructure at xAI and NVIDIA, and its team has tuned kernels that serve billions of tokens a day and designed distributed training systems spanning more than 10,000 GPUs.