Enhancing HPE Generative AI Implementation Service with NVIDIA NIM and NeMo

[HPE_Experts]

HPE Services is excited to announce general availability for our expanded generative AI services capabilities to implement use cases with NVIDIA NIM microservices and NVIDIA NeMo.

By Raffaele Tarantino, WW AI & Data GTM Strategy Lead, HPE AI Services

AI is evolving at unprecedented speed, moving from the top digital transformation agenda topic to a global phenomenon. This shift is leading to the potential next industrial revolution of AI-enabled organizations and AI spearheaders earning prestigious recognition, such as the 2024 Nobel Prize in Physics. Keeping at the same speed of innovation, HPE Services is enhancing HPE Generative AI Implementation Service with the addition of NVIDIA NIM and NeMo, part of the NVIDIA AI Enterprise software platform, now available for organizations engaging with HPE AI Services.

Last year, we launched the HPE Generative AI Implementation Service, and today, we are combining field learnings from customer projects with a close collaboration with NVIDIA. Let’s explore the major benefits to organizations and how we deliver them.

Foundations for a reliable bridge connecting generative AI to industry outcomes

Leveraging on the multi-year experience using full-stack NVIDIA accelerated computing, which includes the NVIDIA AI Enterprise software platform,  as well as the HPE Deployment and Integration for NVIDIA NGC catalog, the HPE AI experts focused on the rich set of NVIDIA NIM microservices available as containers to run anywhere, as well as NVIDIA AI Blueprints that are available for  specific customers’ workflow needs. The use of NVIDIA AI means significant improvements for inference performance, delivering both latest state-of-the-art AI models and zero trade-off on scalability applying dedicated serving profiles, depending on the NVIDIA GPUs in use and related memory available. When needed, HPE AI experts apply NeMo and NVIDIA AI Enterprise components to deliver competitive advantage in the AI Solutions build. Relevant examples of AI solutions include:

• Processing generative virtual screening pipelines on private data, with tailored user experience bridging the technology domain to the scientific research, providing seamless experiences, and unlocking new discovery paths for life science and chemical manufacturing
• Enhancing patient experiences by aggregating heterogeneous data sources, such as medical records and research documents, with advanced RAG with graph methods. Applications for healthcare organizations range from personalized treatment plans proposition to self-service condition status check
• Improving LLM quantitative reasoning for FSI documents review, in activities such as chat, question and answering, and summarization by applying AI agents and securing the vector generation and storing steps. This is extremely important for data confidentiality, and explainable results
• Probing LLM vulnerabilities with dedicated techniques helpful to evaluate the cybersecurity of generative AI adoption. Within this context, NVIDIA NeMo Guardrails plays a key role balancing organizations’ needs and AI output control, fundamental for a trustworthy AI implementation

Be practical and take the time to build generative AI adoption success

Every implementation has its own specifics. Nevertheless, the HPE Generative AI Implementation Service comes with two engagement options moving organizations from Day 0 to GenAI Production-ready application run.

First option offers project management and AI expertise in a three-month engagement providing:

• Project plan – jointly plan the work schedule and refine the requirements, reviewing risks and initiate the project.
• Prompting – refining generative AI model’s input to produce the desired output, without updating the actual weights of the selected model, with techniques like retrieval-augmented generation (RAG), graph topologies, chain of thoughts (CoT), and AI agents.
• Test and deploy – testing and scaling the deployment of generative AI model’s input and output.
• Handover and support – building solution documentation and training for knowledge transfer.

Second option, suitable for deeper customization needs, extends the activities and profiles described above to six-month engagement also providing:

• Fine tuning – refining generative AI model’s input to produce the desired output by directly updating the model parameters using a dataset that captures the knowledge domain you want it to add.

Depending on the complexity of what is taken in scope, HPE AI experts can opt for shorter or longer engagements. The combination of encountered challenges, complexities, benefits, success metrics, and innovation discoveries characterizes the generative AI adoption journey. Balancing all these factors is the key for success.

Moving forward together with HPE AI Services

Reach out to HPE Advisory & Professional Services representatives in your region or plan to visit one of our amazing HPE Customer Innovation Centers across the globe, HPE AI experts will delight you with dedicated technology demonstrations and tailored experience for your priority agenda topics.

---

Meet HPE Blogger Raffaele Tarantino, WW AI & Data GTM Strategy Lead, HPE AI Services

Raffaele Tarantino is the WW AI & Data GTM Strategy Lead for the HPE Advisory & Professional Services. He is responsible for the messaging and sales enablement of the services portfolio. Raffaele has 9 years' experience in HPE roles ranging from private cloud consultant to compute specialist and AI architect as a member of the worldwide practice. Raffaele designed the HPE Machine Learning Development Services and contributed to the HPE Generative AI services launch.

Contact Raffaele on LinkedIn: Raffaele Tarantino

HPE Experts
Hewlett Packard Enterprise

twitter.com/hpe
linkedin.com/company/hewlett-packard-enterprise
hpe.com