RadixArk
AI infrastructure for open access
CORE INFO
RadixArk is a deep-tech company focused on developing large-scale inference and training systems for the AI community. Founded in 2025, the company is headquartered in Sacramento, California, and is committed to making frontier-level AI infrastructure open and accessible to everyone. RadixArk emphasizes first-principles systems design and open infrastructure, contributing significantly to the AI community. The company has developed notable technologies such as SGLang, an open-source inference engine that efficiently serves modern AI models, and Miles, a framework for large-scale reinforcement learning training. RadixArk's solutions are deployed across hundreds of thousands of GPUs, serving trillions of tokens daily for major clients like Google, Microsoft, and xAI. With a team size of 11 to 50 employees, RadixArk continues to innovate in the AI infrastructure space.
WHY WE WOULD WORK AT RADIXARK
Pioneering AI Infrastructure
Join a team dedicated to making frontier-level AI infrastructure open and accessible, revolutionizing how AI models are served globally.
Innovative Tech Stack
Work with cutting-edge technologies like SGLang and Miles, deployed across hundreds of thousands of GPUs for major clients like Google and Microsoft.
Collaborative Team Culture
Be part of a tight-knit team of 11-50 passionate individuals, fostering a collaborative environment that values first-principles systems design.
Strong Financial Backing
Benefit from a $100M seed funding, supported by top investors like Accel and Spark Capital, ensuring stability and growth opportunities.
Impactful Client Partnerships
Contribute to projects serving trillions of tokens daily, impacting major players in the AI industry such as xAI and Microsoft.
Commitment to Open Access
Engage in a mission-driven company that emphasizes open infrastructure, significantly contributing to the AI community's advancement.
MARKET AND TRACTION
GROWTH TACTICS
NOTABLE CUSTOMERS
KEY METRICS
✦ KEY METRICCOMPETITIVE ADVANTAGE
MARKET POSITION
PRODUCT AND TECH
SGLang Inference Engine
An open-source inference engine designed to efficiently serve modern AI models. It is deployed across hundreds of thousands of GPUs, supporting major clients like Google and Microsoft.
Miles Training Framework
An open-source framework for large-scale reinforcement learning training. It empowers AI developers to train models at scale, enhancing learning efficiency and performance.
Open AI Infrastructure
RadixArk's commitment to making frontier-level AI infrastructure open and accessible. This approach democratizes AI technology, enabling broader community engagement and innovation.
First-Principles Systems Design
A foundational approach emphasizing the creation of AI infrastructure from basic principles. This methodology ensures robust, efficient, and scalable solutions for the AI community.
Trillion-Token Scalability
RadixArk's systems are capable of serving trillions of tokens daily, showcasing their infrastructure's scalability and reliability in handling massive data loads.