Skip Navigation
Search

Zeblok - Company Spotlight

 

ZEBLOK

 

News about Artificial Intelligence (AI) is taking the world by storm. With the release of ChatGPT from Open AI Foundation, AI has now become mainstream, with seemingly daily news and information on new AI applications, tools and solutions, but also in the world of banking, art, marketing, entertainment and more. All of these AI based solutions are developed by software development companies just like traditional software, but are based on machine learning, model development, and development of AI inference engines. 

The Zeblok Ai-MicroCloud® is an integrated AI platform-as-a-Service organizations use to create inferences, a method to use AI to process information to make logical deductions and predictions. It is a build-once-run-anywhere abstraction, like what virtualization did to bare metal but for cloud-native architecture. Ai-MicroCloud® also provides workflows for provisioning and orchestrating Artificial Intelligence workloads across public cloud, on-premises and edge servers. The Ai-MicroCloud® is purpose built to enable AI teams to create AI based applications and inferences, and deploy them quickly and efficiently across public clouds, on-premises data centers, and edge computing environments.

Zeblok’s Ai-MicroCloud® was first built and delivered within CEWIT utilizing the Reality Deck NVIDIA RTX6000 cluster. CEWIT staff significantly helped in racking and cabling the servers to support cycle stealing of capacity when the Reality Deck was not utilized. Early adoption by faculty and students in the computer science department and by Laufer Center researchers added a significant impetus to the venture and helped improve the product quality.

Ai-MicroCloud® is an excellent workforce development platform for next generation high tech jobs given its simple user experience and a productive environment for developers based on open standards and open solutions. It focuses on three objectives, that, in its ultimate impact, can be quite dramatic. 

These three objectives are:

  • Embracing simplicity (Zero or no training for developers and modelers with awareness of standard open source tools)
  • Increasing productivity (AI output per unit time), and 
  • Enhancing performance (AI accuracy vs cost trade off) 

Ai-MicroCloud® does this by creating a single, integrated platform that follows three principles:

  • Aggregate ISVs, models, toolkits, AI assets, using existing and new tools and technologies. 
  • Integrate these technologies upon an abstracted infrastructure layer, supporting any topology and combination of chips and hardware.
  • Promulgate by delivering these assets and technologies as software (SaaS), driven by platform services (PaaS), running on any infrastructure (IaaS).

Zeblok is in Intel’s winner circle and has the unique distinction of supporting both NVIDIA and Intel architectures within its Ai-MicroCloud® making it a truly heterogeneous platform. Zeblok looks forward to further helping the ecosystem of startups attached to various incubators within Stony Brook and with faculty/students towards workforce development.

For more information, please visit us at Zeblok.

top