AI-WorkStation

PR2.jpg

Fresh

MARKET

The Zeblok AI-WorkStation provides a single unified environment, which seamlessly orchestrates infrastructure, open source AI/ML frameworks and proven original AI algorithms. An intuitive and simple user interface enables a data scientist, data engineer, or AI engineer to access petaFLOPS of computing power, efficiently develop and train new AI models and take them to production in a runtime environment on the same platform, with consummate ease.

 

The Zeblok AI-WorkStation builds upon and virtualizes the ubiquitous Jupyter Notebook technology, integrating a powerful multi-class orchestration and scheduling layer to support a variety of workloads, from a single GPU to hundreds. Under the hood, our container-based orchestration engine supports both straightforward (non-computationally intensive) AI model development, as well as computationally-intensive workloads.

The Zeblok AI-WorkStation provides enormous freedom to run enterprise workloads in many data centers based on cost profile and tailored to application requirements. This flexibility to drive their workloads is capital-efficient. A user can select low latency compute resources from business partners' data centers or from (albeit higher priced) public clouds like AWS, GCP and Azure.

Data scientists don't have to waste time optimizing frameworks on GPU platforms. CUDA optimization for popular AI frameworks is available out of the box and ready to go with the click of a button. Popular data science language bindings, such as R, Scala, and Python are also available. We leveraged a community of data scientists to beta-test and harden these containers so enterprise developers can spend their time developing innovative models, rather than attending to infrastructure.

 

The Zeblok AI-WorkStation also makes it easy for developers to build and share new containers with fellow team members, improving organizational productivity. Dependency management, which is involved in AI-development, is greatly simplified and incorporated with examples for a developer to implement high-quality software. Zeblok's AI-WorkStation also makes it a breeze to promote models developed into APIs through a few simple commands.

Go ahead and confidently access petascale capacity through a few clicks from any laptop.

Simple User Interface

The Zeblok AI-Workstation has an extremely simple user interface turns any web browser into a supercomputer 

Familiar Data Science Tools

Access to ubiquitous data science tools like Jupyter Notebooks, Docker, Python, R, Scala make it a zero-learning-curve workstation for data scientists.

Supercomputing - Enormous GPU Capacity

Create and distribute jobs across a network of GPUs, allowing cost-effective and speedy development of hyper-converged-AI applications. Start in minutes through a web browser and self-registration, with 24 PFLOPS tensors (Machine Learning), expanding to 200+ Petaflops.

Novel AI Algorithms

  • Access to a carefully selected library of proven original AI Algorithms, ready to be integrated into enterprise business processes

  • Customizable to your customers' specific needs or the demands of specific business processes within your workflow

  • Scalable instantly

  • Secure, with our VPN, to maintain the integrity of your processes' intellectual property

Multi-User, Multi-Class Orchestration

Let all your data scientists and engineers run workloads into one platform. Orchestration provides Kubernetes-as-a-service and Slurm-as-a-service.

Zeblok logo original-01 (1).jpg
LINKS
ABOUT

Email: zeblok@zeblok.com

Tel: +1 (631) 223-8233

HQ Office:

1500 Stony Brook Road

Stony Brook, NY 11794

Office:

51 JFK Parkway

First Floor West

Short Hills, NJ 07078

    © 2020 Zeblok Computational, Inc.