Build and Scale GenAI Development Agents Securely with Ona and Amazon Bedrock on AWS

By: Len Gomes, Partner Solutions Architect – AWS 
By: Angelo Malatacca, AWS Partner Solutions Architect – AWS
By: Daniel Zäh, Global Partnerships Lead – Ona 
By: Will McMullen, Product Enablement Lead – Ona

Ona

Building brand new Generative AI (GenAI) applications or infusing new GenAI based features into existing ones, still requires maintaining full governance and security boundaries of both the application and the development platform. New GenAI based tools and agents help accelerate the development pace, but they do not remove the need to maintain a fully compliant security posture. In this blog post, we cover how AWS and Ona help you build and scale GenAI-native development environments with agents and guardrails; from initial experimentation all the way to secure, compliant, production-ready systems.

Enterprises must address security, compliance, and governance requirements over the full design, development and deployment cycle of GenAI features while maintaining agility. Regardless of whether you are exploring your first steps into AI-assisted autocomplete, agentic development with Amazon Q Developer, or automating repetitive tasks with an ADE (agentic development environment), you need to ensure that:

• Customer data, source code, and secrets are properly protected.
• Your security perimeter does not expand unnecessarily due to the use of third-party tools that are not compliant with your security policy.
• You maintain a high-speed pace of innovation, with flexibility to adopt the next AI breakthrough and frontier models or tools as they are released.

In addition, enterprise use cases require HIPAA (Health Insurance Portability and Accountability Act), SOC2 (System and Organization Controls), or FedRAMP (Federal Risk and Authorization Management Program) compliance of your application, and certainly you need to protect your IP (intellectual property). You will also be needing to implement proper guardrails against potential hallucinations made by your LLM (Large Language Model) of choice.

Ona is built for enterprise use cases, to confidently improve access to controlled systems for global and rapidly scaling workforces. The same technology, coupled with a streamlined IDE (Integrated Development Environment) tailored for coding agents, enables customers to roll out GenAI initiatives from steps in autocomplete mode all the way to complex pipelines through parallel, autonomous agent fleets.

Even with a secure foundation for development environments, enterprises still needs a way to secure LLM inference pipelines and host foundational GenAI models. Amazon Bedrock is a fully managed service that provides a unified API to access high-performing foundation GenAI models. Amazon Bedrock offers access to a diverse and expanding portfolio of foundation models from industry-leading AI companies, including Anthropic, Cohere, Meta, Mistral AI, AI21 Labs, Stability AI, OpenAI, and Amazon’s own Nova family of models. It offers a robust set of tools and features designed to help you build generative AI applications efficiently while adhering to best practices in security, privacy, and responsible GenAI. In addition, AWS PrivateLink is available to establish private connectivity from your Amazon VPC to Amazon Bedrock, avoiding potential accidental exposure of customer data to the internet. This approach gives you secure access to GenAI capabilities removing the heavy lifting of managing the underlying infrastructure.

Amazon Bedrock and Ona integration has been designed from the start to offer a secure, standardized, and governed environment to accelerate your adoption of GenAI-powered development technology.

Amazon Bedrock provides comprehensive security controls for managing your generative AI workloads. It integrates with AWS security and governance tools including AWS Identity and Access Management (IAM) for access control, VPC endpoints for network isolation, AWS Key Management Service (AWS KMS) for encryption and AWS CloudTrail for audit logging. Customer data processed by Amazon Bedrock is encrypted in transit and at rest within the AWS region where you use the service, and AWS does not use customer data to train foundation models. The service supports monitoring through AWS CloudWatch and Amazon EventBridge, helping organizations maintain security and compliance requirements.

Ona: Developer freedom without the risk

Ona enables secured, scalable, and standardized development environments built and orchestrated entirely within your Amazon VPC. Every workspace your developers launch is a fresh, temporary Amazon EC2 instance, pre-configured and pre-connected to your company’s essential developer tools and internal services . These are not cloud-hosted environments run by someone else, but rather they are fully contained, short-lived VMs, automatically secured your existing security controls.

Figure 1: Ona’s professional software engineering agents that execute tasks in parallel

Why does that matter? Because developers naturally start AI projects by experimenting, tweaking, and iterating. If these early experiments happen on personal laptops or untrusted Integrated Development Environment (IDEs), you lose visibility and control immediately.

This means developers will have access to AI coding agents running the latest approved models, with the security and access controls your company requires, and the LLM inference providers you are partnered with. All the while, eliminating the complexities and risks associated with managing foundational model access and agentic infrastructure on their own. This solution offers:

• Engineers to code with agentic AI without changing how or where they work.
• Multiplied engineering velocity with AI without losing control of how or where agentic reasoning happens.
• Automatically enable every engineer with highly capable, fully provisioned and pre-configured agents, with no tooling or development environment setup.
• 10x engineering output without adding headcount.

Ona keeps experimentation agile and controlled. Your teams will be able to innovate with agility, yet every action stays auditable, secured, and aligned with your internal policies.

Figure 2: Ona’s Core Building Blocks

Secure Development and Inference for your AI journey

Pairing Ona and Amazon Bedrock work together to improve development velocity while creating a seamless, fully secure workflow for AI from inception to production enables:

• Developers to build and test AI code directly inside secure Ona environments.
• Workspaces that automatically include the required AWS credentials provided securely via OIDC tokens (OpenID Connect), scoped precisely for Amazon Bedrock APIs and other internal AWS services (Amazon Elastic Container Registry, AWS Simple Storage Service, Amazon CloudWatch, etc.)
• Using Amazon Bedrock centrally, while enforcing model permissions, auditing API calls, and tracking usage while ensuring safe, traceable, compliant interactions.

This empowers your entire team with highly governed, AI-native development tools without expanding your security surface, introducing platform complexity, or risking vendor lock-in. AWS Bedrock and Ona together provide a straightforward path to scale AI confidently, even in highly regulated environments.

Why this makes a difference for your team

For platform security teams, this integrated stack allows to maintain high speed of innovation while staying compliant and secure by implementing your security policies. You will be able to build compliant workspace templates, approve specific GenAI models for usage, and manage secrets centrally.

Some of the benefits for platform security teams are:

• GenAI model inference API calls only come only from trusted, authorized VPCs.
• Secrets are dynamically managed via IAM, never exposed or stored in plain text.
• Source code and sensitive data stay entirely within the secure boundaries you define.

Compliance gets simpler: Amazon Bedrock is in scope for common compliance standards such as FedRAMP Moderate, SOC, International Organization for Standardization (ISO), HIPAA eligibility, and customers can use Amazon Bedrock in compliance with the General Data Protection Regulation (GDPR). This service is included in the scope of the SOC 1, 2, 3 reports, allowing customers to gain insights into our security controls. We demonstrate compliance through extensive third-party audits of our AWS controls. Amazon Bedrock is one of the AWS services under ISO Compliance for the ISO 9001, ISO 27001, ISO 27017, ISO 27018, ISO 27701, ISO 22301, and ISO 20000 standards. Additionally, it is CSA Security Trust Assurance and Risk (STAR) Level 2 certified, which validates the use of best practices and the security posture of AWS cloud offerings

Organizations operating in heavily regulated spaces are already seeing tangible benefits by measuring faster onboarding for developers and data scientists, reduction in administrative overhead from managing disparate tools, and substantial decreases in security vulnerabilities from credential exposure and shadow IT. Furthermore, these companies are accelerating the secure development of internal AI capabilities, including custom agents, copilots, and automated workflows, all while maintaining compliance and reducing risk.

Conclusion

Leveraging Amazon Bedrock and Ona creates a powerful solution for organizations implementing AI-native development while maintaining strict security and compliance standards. Amazon Bedrock serves as your secure gateway to GenAI foundation models, allowing organizations to maintain strict governance within their AWS infrastructure. Ona complements this by providing developers with secure, ephemeral AI agents & workspaces directly within your VPC. This creates a seamless development experience where innovation happens in controlled environments that automatically inherit your security policies. The integration eliminates the traditional friction between developer agility and enterprise requirements.

The combination of both, facilitates an end-to-end secure GenAI development pipeline: security teams gain confidence knowing security is robust, compliance teams benefit from Bedrock’s extensive certifications. Moreover, development teams can focus on innovation without wrestling with security constraints. Together, they provide a scalable, secure foundation that enables organizations to confidently build the next generation of GenAI-powered applications while keeping sensitive data and intellectual property protected.

Start building with Ona. check Ona docs, or contact them to explore how Ona can accelerate your engineering teams. Ona is also available on the AWS Marketplace.

Ona – AWS Partner Spotlight

Ona is an AWS Advanced Technology Partner and AWS Competency Partner that provides secure, ephemeral development environments and professional AI agents that collaborate in parallel or autonomously across the software lifecycle. Contact Ona | Partner Overview | AWS Marketplace