YS
Ying Sheng
Co-Founder & CEO
Previously Inference Engineer at xAI
AI infrastructure researcher and engineer who created SGLang at UC Berkeley's LMSYS group in 2023, then built inference systems at xAI before co-founding RadixArk.
RadixArk is an AI infrastructure company that commercializes SGLang, the open-source inference engine deployed across hundreds of thousands of GPUs worldwide and generating trillions of tokens daily for customers including Google, Microsoft, NVIDIA, xAI, and LinkedIn. The company builds an end-to-end AI infrastructure platform covering inference, training, and post-training, with SGLang for inference and Miles for reinforcement learning as its open-source foundations. RadixArk offers managed infrastructure and tooling for developers, enterprises, and research labs to build advanced AI systems with high speed and control.
RadixArk is an AI infrastructure company that commercializes SGLang, the open-source inference engine deployed across hundreds of thousands of GPUs worldwide and generating trillions of tokens daily for customers including Google, Microsoft, NVIDIA, xAI, and LinkedIn. The company builds an end-to-end AI infrastructure platform covering inference, training, and post-training, with SGLang for inference and Miles for reinforcement learning as its open-source foundations. RadixArk offers managed infrastructure and tooling for developers, enterprises, and research labs to build advanced AI systems with high speed and control.
RadixArk traces its roots to 2023 when Ying Sheng and collaborators at UC Berkeley's LMSYS research group (created by researchers from Stanford, Berkeley, CMU, UCSD) built SGLang, an open-source LLM inference engine. Sheng subsequently joined xAI to build inference systems for Grok, while Banghua Zhu worked at NVIDIA. In early 2026, Sheng and Zhu spun RadixArk out of the SGLang project to commercialize the technology, securing $100M in seed funding from Accel, Spark Capital, and NVIDIA at a $400M valuation.
SGLang, the foundation of RadixArk's platform, is deployed across hundreds of thousands of GPUs and generates trillions of tokens daily for customers including Google, Microsoft, NVIDIA, xAI, Oracle, and LinkedIn.
RadixArk launched publicly with $100M in seed funding led by Accel and Spark Capital, with participation from NVIDIA, AMD, Databricks, and angels including OpenAI co-founder John Schulman and PyTorch creator Soumith Chintala.
TechCrunch reported that the SGLang open-source inference project was spinning out as RadixArk, backed by Accel and Spark Capital, as AI inference infrastructure spending accelerated.
Co-Founder & CEO
Previously Inference Engineer at xAI
AI infrastructure researcher and engineer who created SGLang at UC Berkeley's LMSYS group in 2023, then built inference systems at xAI before co-founding RadixArk.
$100M raised total
No H1B sponsorship data available for this company.