CIQ, the founding support and services partner of Rocky Linux, is launching Service Endpoints, a new capability for its Fuzzball platform that enables Fuzzball to be a turnkey, sovereign AI infrastructure platform by unifying model training, fine-tuning, validation, and inference in single, portable workflows.
According to the company, organizations can now develop and serve AI models entirely from on-premises or hybrid environments, eliminating reliance on external platforms while maintaining complete control over proprietary and sensitive data.
Service Endpoints brings training and inference into a single, portable workflow so teams stop treating deployment as a separate platform and a manual handoff.
Service Endpoints enables users to define and run persistent endpoints alongside training and fine-tuning steps within the same orchestrated pipeline. The result is accelerated time to first token, fewer brittle deployment pipelines, less unbudgeted technical debt, and more time reclaimed for iteration and innovation as AI initiatives scale, the company said.
The capability combines batch computing (for training and fine-tuning) with persistent services (for high-performance inference and interactive development) in unified workflow definitions.
"Organizations want to use AI with their proprietary data in environments that they control, without sending their data to external platforms," said Jonathon Anderson, Fuzzball product lead at CIQ. "Fuzzball Service Endpoints treats the entire AI stack as a composable unified workflow that can be iterated on and refactored to meet your specific requirements. This gives you complete control over your AI experience, whether you’re fine-tuning a model or orchestrating a suite of coordinating agents."
Service Endpoints provides native support for Jupyter, VDI, visualization, and other services while enabling real-time inspection and adjustment of running workflows. This allows researchers to observe simulations in progress, validate results midstream, and guide computations as they run, accelerating innovation cycles across research domains, the company said.
The new capability enables three distinct categories of high-impact use cases:
- Turnkey sovereign AI
- High-performance service (inference) processes
- Interactive access to HPC resources
Fuzzball Service Endpoints is designed specifically for performance-sensitive environments. It preserves bare metal performance while introducing service-oriented flexibility, CIQ said.
For more information about this news, visit https://ciq.com.