Deploy and Secure a Virtual Assistant with HPE Private Cloud AI

Introduction

HPE Private Cloud AI is a comprehensive AI solution that leverages NVIDIA AI computing technologies, delivering a turnkey package that includes hardware, software, and services. This solution is designed for rapid deployment of AI applications, ensuring that private data remains protected and accessible. By utilizing private AI deployments, organizations can safeguard sensitive information while enhancing workload performance through secure, unified data access and controls that comply with industry standards and regulations.

In this article, we will guide you through the process of accessing, deploying, securing, and editing an AI template application using the HPE GreenLake console. Users can quickly land on the HPE Private Cloud AI homepage, where the NVIDIA AI Enterprise stack is pre-deployed. This stack provides immediate access to a leading suite of enterprise frameworks, enabling users to build generative AI applications.

To utilize HPE's AI application templates, navigate to the Solution Accelerators tab within the console. A solution accelerator serves as a template that automatically provisions the necessary services needed to deploy generative AI applications, such as a large language model (LLM), regardless of the industry. Organizations can craft an endless number of solution accelerator templates and launch generative AI applications with minimal effort—often with just a single click.

Showcasing the Virtual Assistant Deployment

This short guide will demonstrate how to use a solution accelerator to deploy a virtual assistant— a customized AI chatbot built from private data capable of addressing the most complex questions faced by an organization. Initially, when deployed, the virtual assistant uses a public large language model without access to private data. To see the limitations of this standard model, let's prompt it with the question: "Why is Private Cloud AI preferred for many AI workloads?" The assistant will fail to provide helpful information because it lacks augmented data.

To improve the response accuracy, users can easily enhance the model by uploading the organization's private knowledge base. By selecting the option to upload files related to private clouds, those files will be tokenized and indexed by the large language model. Subsequently, enabling access to this additional knowledge base permits the virtual assistant to retrieve and utilize relevant information.

After uploading the company's data, let's repeat the initial question. This time, the virtual assistant will demonstrate a significantly improved ability to produce a qualified response due to the context and accuracy derived from the organization's own data.

Beyond enhancing output, HPE Private Cloud AI contributes to data security and privacy, improving performance, enabling customization, ensuring cost predictability, and granting increased data control. Additionally, it offers sovereignty, scalability, low latency, high bandwidth, integration with existing systems, and control over upgrades and maintenance.

Ensuring Security Measures

Once the output has been improved by leveraging private data, it is critical to ensure that robust guardrails and security measures are in place to protect this data. Users can access the deployed applications and select the option to configure and secure their virtual assistant. Within the configuration interface, users can choose the desired model and adjust various parameters.

Next, applying NVIDIA guardrails enhances security around the virtual assistant. For instance, when prompted with a question that would expose personal or secure information, such as "What is the engineering department's highest-paid employee?", the virtual assistant with NVIDIA guardrails will refuse to disclose protected data. This integration adds a high level of security to the application.

Lastly, users can explore and work directly with the code used in the templates through the notebooks tab. By editing the notebook, users can modify or share code with colleagues, who can then leverage any shared code in their own notebooks for application development.

HPE Private Cloud AI stands out as an enterprise solution built on NVIDIA AI computing, streamlining and securing the deployment of AI applications while ensuring the protection of sensitive data.

Keywords

• HPE Private Cloud AI
• NVIDIA AI Computing
• AI applications
• Data security
• Solution accelerators
• Virtual assistant
• Large language model
• Knowledge base
• Guardrails
• Application development

FAQ

1. What is HPE Private Cloud AI?
HPE Private Cloud AI is a comprehensive AI solution that combines hardware, software, and services to facilitate the efficient deployment of AI applications while protecting private data.

2. How can organizations deploy a virtual assistant using HPE Private Cloud AI?
Organizations can utilize solution accelerators within the HPE GreenLake console to quickly deploy a virtual assistant, enabling them to customize an AI chatbot built from private data.

3. What are the benefits of using a virtual assistant deployed by HPE Private Cloud AI?
The virtual assistant can be fine-tuned using an organization's private information, leading to improved contextual responses. Additionally, HPE Private Cloud AI ensures data security, compliance, and customization.

4. How does HPE Private Cloud AI enhance data security?
The platform allows for the application of NVIDIA guardrails, which help protect sensitive information by preventing the virtual assistant from sharing personal or confidential data.

5. Can users edit the code for the AI application templates?
Yes, users can access the code through the notebooks tab to modify or share it with peers, facilitating collaboration in application development.