Ai-MicroCloud®
Ai Platform-as-a-Service
Powering Business Innovation
Portability
- AWS, Azure, GCP in under 15 min
- On-premise in under 30 min
- Multi-cloud
- Cloud-2-edge
- Build-once-run-ai-anywhere
Productivity
- Modelers launch workstation in seconds
- Ops deploy services in minutes
- Ops model serve vision, LLM, NLP, and other models in minutes
- Support for heterogeneous chips
Composability
- Plug n play third party tools in minutes
- Plug n play for HPC and addon orchestration
- Ready to consume foundational models
Scalability
- Uniquely combine ML Ops & Infra
- Scale up for model fine-tuning
- Scale up inference deployment from Cloud to thousands of Edge locations
Collaboration & Reuse
- Product, Modeler, Ops roles
- Aggregate Ai assets for reuse
- Snap shot training environment
- Automate deployment
- Rapidly build and deploy GenAi, ML, and DL applications
Enterprise Integration
- With enterprise ML Ops
- With enterprise Dev Sec Ops
- With developer tool chains
- Python SDK for automation
- CLI for model serving
Architecture Diagram
Composable Foundational Components
SaaS Components
Ai-WorkStation & Ai-HPC-WorkStation
A productive web-based developer environment for Ai model trainers with dynamically provisioned resources. Best way to rapidly create end-to-end Ai pipelines on heterogeneous chip architectures.
Data Science
A versatile web-based digital asset repository for ready-to-use Ai inference engines from various sources. Can launch pre-existing Ai-model training environments and inference engines.
Product Management
A scalable web-based ML Ops env. for packaging and publishing Ai inference engines as a secure network asset to Edge. Enabling deployment, providing a consistent publishing mechanism.
PaaS Components
Microservice Manager
Microservice Manager efficiently delivers Ai applications seamlessly across hybrid-cloud, multi-cloud environments, and Edge data centers in compliance with customer policies
Ai-WorkStation Manager
Ai-WorkStation Manager enables zero learning curve, open-source Ai frameworks, curated Ai applications, other end-to-end Ai pipelines as virtualized notebooks while increasing shareability
Ai-HPC Manager
Ai-HPC Manager seamlessly scales to high-performance computing (HPC), crossing physical server boundaries for Ai model training and optimization, using familiar and consistent user experience
Ai-Database Manager
Ai-Data Lake Manager improves productivity and data handling for model training and inference engine optimization, while accelerating data manipulation tasks when GPUs are available within the underlying Ai-MicroCloud® environment
IaaS Components
RBAC Manager
Role-based access control (RBAC) restricts network access based on a person's role within an organization and/or Ai project team. It enables advanced access control for all SaaS and PaaS application components within the platform
Security Manager
A comprehensive security capability maps users seamlessly and securely between web-domain and meta-scheduled infrastructure across hybrid-cloud environments. Integration with enterprise LDAP and/or Active Directory makes it easy to bring consistent user and group policies, including single sign-on and multi-factor authentication.
Multi-cloud Orchestration
A scalable web-based ML Ops env. for packaging and publishing Ai inference engines as a secure network asset to Edge. Enabling deployment, providing a consistent publishing mechanism.
©️ Zeblok Computational Inc. 2022