AWS Marketplace

Streamline scalable AI governance with Domino in AWS Marketplace

Enterprise AI adopters are setting big innovation goals that will demand more from their governance capabilities. Domino Data Lab, an AWS Partner and seller in AWS Marketplace, provides its Enterprise AI Platform to help organizations balance speed of innovation with governance requirements. In AWS Marketplace, teams can quickly obtain and deploy Domino’s platform on Amazon Web Services (AWS), enabling rapid AI transformation and machine learning operations (MLOps) while maintaining compliance and cost efficiency through optimized AI infrastructure. The platform works with AWS specialized AI processors—AWS Inferentia for inference workloads and AWS Trainium for accelerated training—to provide a solution that speeds up AI implementation.

This post examines how Domino’s Enterprise AI Platform and AWS machine learning (ML) accelerators help organizations innovate with AI while preserving governance and performance. We also explore how Domino simplifies deployment to Amazon SageMaker, helping teams accelerate AI development responsibly.

Balancing AI and governance

AI governance goes beyond compliance—it requires promoting transparency, repeatability, and risk management without slowing down model development. Organizations scaling AI face several common challenges:

  • Implementing risk management and compliance requirements while maintaining innovation speed
  • Maintaining clear model lineage and approval workflows across teams
  • Establishing consistent deployment policies across environments

Organizations need a structured approach to AI governance that implements controls while preserving agility. Through automated policies, approval workflows, and compliance tracking, teams can maintain oversight without creating friction.

Scalable, governed AI development with Domino and AWS

To scale AI effectively, teams need flexible infrastructure that maintains governance standards. Domino’s Enterprise AI Platform give data scientists the flexibility to use their preferred tools (including TensorFlow, PyTorch, R, or SAS) and train models on AWS compute resources while tracking provenance and enforcing practices automatically. Models developed in Domino can be deployed to Amazon SageMaker for running on AWS hardware, maintaining consistent governance throughout.

AI development requires collaboration between data scientists, IT teams, and compliance specialists. Domino supports this through version control, shared workspaces, and model lineage tracking. The platform helps IT and risk teams implement compliance controls early in development by defining policies, managing approvals, and automating audits.

Enhancing AI performance with Inferentia and Trainium

Deploying AI models at scale creates specific challenges. Organizations need to manage deployment workflows across hybrid environments while controlling costs. As AI models require more computing power, inference costs can increase without optimized hardware. Additionally, maintaining governance and compliance across distributed systems is important to prevent model drift and meet regulatory requirements.

AWS addresses these challenges with specialized AI chips that Domino uses for enterprise AI workloads:

  • AWS Inferentia delivers high-performance, low-cost inference optimized for deep learning models, reducing the compute burden of running large-scale AI applications.
  • AWS Trainium accelerates training workloads, improving efficiency and cost-effectiveness for AI development.

Domino integrates with these AWS processors while maintaining governance and lifecycle tracking. Models trained on AWS Trainium can be deployed to Inferentia-based Amazon SageMaker endpoints through Domino’s platform. The platform tracks all model deployments, allowing IT teams to scale AI services while managing costs and compliance requirements.

Model deployment to Amazon SageMaker

Domino offers integration for deploying models to Amazon SageMaker. This integration helps MLOps engineers and data scientists deploy models while following governance requirements. Here’s how it works:

Step 1: Configuring deployment targets

To ensure seamless deployment, Domino administrators configure Amazon SageMaker as an external deployment target within the Domino platform. This setup allows data scientists and MLOps teams to deploy models efficiently while maintaining centralized governance.

Figure 1: Domino platform’s admin view of creating Amazon SageMaker external deployment targets

Figure 1: Domino platform’s admin view of creating Amazon SageMaker external deployment targets

Organizations can use AWS Inferentia-backed instances to optimize inference costs and performance. This configuration ensures that organizations achieve high efficiency at scale without excessive infrastructure costs.

Figure 2: Domino platform’s admin view of configuring AWS Inferentia resources and permissions

Figure 2: Domino platform’s admin view of configuring AWS Inferentia resources and permissions

Step 2: Deploying models

After the deployment target is configured, data scientists can select trained models within Domino’s UI and package them for deployment to SageMaker. The integration enables a streamlined process where models are automatically containerized and optimized for AWS Inferentia.

After packaging, an inference endpoint is created within SageMaker, allowing for low-latency, high-performance AI inference. This automation significantly reduces the time and effort required to move models from development to production.

Figure 3: Select the model that will be deployed to Amazon SageMaker

Figure 3: Select the model that will be deployed to Amazon SageMaker

Figure 4: Select deployment target and size (previously configured by Domino Admin)

Figure 4: Select deployment target and size (previously configured by Domino Admin)

Figure 5: Configure endpoint parameters, such as instance count, streaming, and auto scaling

Figure 5: Configure endpoint parameters, such as instance count, streaming, and auto scaling

The full demonstration video of Domino seamless deployment to Amazon SageMaker is available on YouTube at Deploy to Amazon SageMaker Seamlessly with Domino.

Step 3: Managing and monitoring AI workloads

After deployment, real-time monitoring and performance tracking become crucial. Domino provides a centralized interface to track deployed models, monitor inference latency, and promote compliance with governance policies.

With built-in model lineage tracking and automated reporting, IT and compliance teams can oversee inference endpoints across environments, reducing the risk of model drift or performance degradation.

Figure 6: Manage and oversee all assets, regardless of where they’re produced or deployed, through Domino’s unified, single pane of glass

Figure 6: Manage and oversee all assets, regardless of where they’re produced or deployed, through Domino’s unified, single pane of glass

Use cases and benefits

The following sections explore the use cases and benefits of Domino’s solution.

Automated CI/CD for AI models

Modern AI applications require continuous iteration, updates, and improvements. However, managing these changes manually can be time-consuming and error-prone. Domino automates the continuous integration and continuous delivery (CI/CD) process for AI models, which means that updates, retraining, and redeployments happen efficiently and with full version control. This automation enables:

  • Streamlined model packaging and deployment through integrated workflows
  • Automatic tracking of model changes, providing reproducibility and auditability
  • Policy-based approvals to enforce governance while allowing teams to iterate quickly

By embedding CI/CD best practices into model development, enterprises can accelerate AI innovation while maintaining robust oversight.

Infrastructure optimization with AWS compute resources

Optimizing AI infrastructure costs while maintaining performance is a key concern for enterprises scaling AI workloads. Domino provides flexibility to take advantage of the full spectrum of AWS compute resources. AWS Trainium accelerates model training, reducing training time and compute costs. AWS Inferentia delivers optimized inference performance, reducing operational expenses for large-scale deployments. For workloads requiring NVIDIA acceleration, Domino supports various Amazon Elastic Compute Cloud (Amazon EC2) GPU instance types (such as P4, G5, and G4) to handle demanding deep learning tasks. Domino enhances these capabilities by:

  • Enabling auto scaling of inference workloads to dynamically adjust capacity based on demand, preventing overprovisioning
  • Supporting cost-efficient model deployment across different compute types, allowing organizations to match resources to workload requirements
  • Scaling down to zero when models aren’t actively in use, reducing unnecessary cloud expenses

These optimizations mean that enterprises can achieve the best possible balance between cost-efficiency and AI performance, regardless of which AWS compute resources best suit their specific needs.

Who benefits?

Implementing AI with Domino through AWS Marketplace provides specific value to different organizational roles:

  • Data science teams – On-demand scalability means that data science teams can access AWS Inferentia and Trainium through Domino for faster experimentation without IT bottlenecks. Collaboration across teams is seamless, meaning they can share and reuse projects in a centralized, version-controlled workspace. Best practices are built in so they can automatically track experiments and provide reproducibility and documentation.
  • MLOps engineers and risk and compliance teams – Deployment is simplified, so engineers can push models to Amazon SageMaker in a few clicks, with built-in control over scaling and instance types. Centralized monitoring means being able to manage all model endpoints—on-premises, cloud, or SageMaker—from one interface. Teams are compliance-ready with full audit trails and policy enforcement to meet internal and regulatory standards.
  • IT leaders and data science executives – Reduce infrastructure setup time with a ready-to-use, cloud-native platform for faster time to value. Improve cost-efficiency by using AWS Marketplace for streamlined procurement and cost control across AI workloads. Reduce risk with standardized governance, improved model visibility, and fewer shadow AI deployments.

Conclusion

Scaling AI within an enterprise requires a careful balance of performance, governance, and cost-efficiency. Together, Domino Data Lab and AWS strike that balance, enabling successful AI modernization while maintaining robust model risk management and enterprise integration. AWS Inferentia and Trainium provide the high-performance, low-cost compute foundation for training and deploying modern AI models, while Domino’s platform ties it all together with seamless workflows, collaboration, and end-to-end governance.

Next steps:

About Authors

James Yi

James Yi is a is a Senior AI/ML Partner Solutions Architect at AWS. He spearheads AWS’s strategic partnerships in Emerging Technologies, guiding engineering teams to design and develop cutting-edge joint solutions in generative AI. He enables field and technical teams to seamlessly deploy, operate, secure, and integrate partner solutions on AWS. James collaborates closely with business leaders to define and execute joint Go-To-Market strategies, driving cloud-based business growth. Outside of work, he enjoys playing soccer, traveling, and spending time with his family.

David Schulman

David Schulman is a data and analytics ecosystem enthusiast in Seattle, WA. As Director of Global Partner Marketing at Domino Data Lab, he works closely with other industry-leading ecosystem partners on joint solution development and go-to-market efforts. Prior to Domino, David lead Technology Partner marketing at Tableau, and spent years as a consultant defining partner program strategy and execution for clients around the world.