Platform · Cloud & Pods
Use NeuroCluster Cloud for model training and additional compute capacity. Deploy NeuroCluster Pods as a private cloud inside your organisation for local development and processing data that cannot leave your organisation.
Deployment scope, capacity and data boundaries are agreed around your workload.
Two deployment environments
Training and development have different demands. Cloud provides a place for compute-intensive work; Pods bring the platform into the environment where your teams and protected data need to work.
Training & additional capacity
Give demanding workloads room to run.
Use cloud compute for model training and work that needs additional capacity. Agree the resources, location and information permitted for each workload before it runs.
Cloud work uses the data and workloads approved for that environment.
Your organisation’s private cloud
Keep protected work close to your teams.
Deploy Pods as a private cloud within your organisation’s own environment. Develop locally and process protected data inside that boundary, with access and operating responsibilities defined for the deployment.
Data that cannot leave your organisation is processed locally in Pods.
At a glance
Choose the environment for the workload. The deployment design sets out what runs where, which information it may use and who is responsible.
Swipe across to compare both environments.
| Requirement | NeuroCluster Cloud | NeuroCluster Pods |
|---|---|---|
| Model training | Train models using approved datasets and agreed compute resources. | Develop and validate locally; agree any local training requirements in the deployment scope. |
| Compute capacity | Additional capacity for workloads approved to run in Cloud. | Compute within your organisation’s environment or private cloud, sized for the local workload. |
| Local development | Cloud work can support the training and compute needs of a development project. | A local environment for developing and validating applications and workflows. |
| Data boundary | Only the information and workloads approved for the Cloud environment. | Protected data processing remains within the chosen Pods boundary. |
| Location | NeuroCluster’s European managed cloud environment. | A private cloud deployed inside your organisation. |
| Working together | Runs the training or additional compute included in an approved hybrid design. | Runs local development and protected processing; Cloud work is separately scoped and approved. |
The final design covers data access, permitted transfers, model use and operational responsibilities for each environment.
A practical hybrid approach
A team develops an application in Pods using protected operational data. If the project needs model training or more compute, first define a separate Cloud workload and the information it may use. Restricted data stays in the Pods workflow.
01
Identify the restricted data, the work that must run in Pods and the people responsible for access and approval.
02
Run local development and protected data processing inside your organisation or private cloud under the agreed controls.
03
Scope the training or additional compute, review its input data and authorise only the work permitted in Cloud.
04
Review the resulting models or outputs and agree how they may be used in the application before operational rollout.
Illustrative deployment approach. Data movement and any model or result handoff must be explicitly designed and approved for the use case.
Deployment and products
NeuroCluster Cloud and Pods describe deployment environments. Agentic Cloud is the software product that coordinates supported agents around shared context and defined authority. Its name does not determine where your workloads must run.
Coordinate supported agents, ownership and approved actions. Agree its deployment around the workflow’s requirements.
Apply configured access and policy controls to AI requests routed through the gateway, with destinations agreed for the deployment.
Shape scoped applications and reporting workflows on the shared foundation. Compose is in development preview.
Plan the full deployment
Review the platform, private deployment options and trust material alongside your workload requirements.
Start with your workload
Bring the training or development goal, the data that needs protection and the capacity you expect. We will discuss the right Cloud, Pods or hybrid scope.
