Covering Disruptive Technology Powering Business in The Digital Age

NVIDIA has announced AWS re:Invent a generative AI (artificial intelligence) microservice that lets enterprises connect custom large language models to enterprise data to deliver highly accurate responses for their AI applications.

NVIDIA NeMo™ Retriever—a new offering in the NVIDIA NeMo family of frameworks and tools for building, customizing, and deploying generative AI models—helps organisations enhance their generative AI applications with enterprise-grade retrieval-augmented generation (RAG) capabilities.

As a semantic-retrieval microservice, NeMo Retriever helps generative AI applications provide more accurate responses through NVIDIA-optimised algorithms. Developers using the microservice can connect their AI applications to business data wherever it resides across clouds and data centres. It adds NVIDIA-optimised RAG capabilities to AI foundries and is part of the NVIDIA AI Enterprise software platform, available in AWS Marketplace.

Cadence, Dropbox, SAP and ServiceNow are among the pioneers working with NVIDIA to build production-ready RAG capabilities into their custom generative AI applications and services.

“Generative AI applications with RAG capabilities are the next killer app of the enterprise,” said Jensen Huang, Founder of and CEO at NVIDIA. “With NVIDIA NeMo Retriever, developers can create customised generative AI chatbots, copilots, and summarization tools that can access their business data to transform productivity with accurate and valuable generative AI intelligence.”

Global Leaders Enhance LLM Accuracy with NeMo Retriever

Electronic systems design leader Cadence serves companies across hyperscale computing, 5G communications, automotive, mobile, aerospace, consumer, and healthcare markets. It is working with NVIDIA to develop RAG features for generative AI applications in industrial electronics design.

“Generative AI introduces innovative approaches to address customer needs, such as tools to uncover potential flaws early in the design process,” said Anirudh Devgan, President and CEO at Cadence. “Our researchers are working with NVIDIA to use NeMo Retriever to further boost the accuracy and relevance of generative AI applications to reveal issues and help customers get high-quality products to market faster.”

Cracking the Code for Accurate Generative AI Applications

Unlike open-source RAG toolkits, NeMo Retriever supports production-ready generative AI with commercially viable models, API stability, security patches and enterprise support.

NVIDIA-optimised algorithms power the highest accuracy results in Retriever’s embedding models. The optimised embedding models capture relationships between words, enabling LLMs to process and analyse textual data.

Using NeMo Retriever, enterprises can connect their LLMs to multiple data sources and knowledge bases, so that users can easily interact with data and receive accurate, up-to-date answers using simple, conversational prompts. Businesses using Retriever-powered applications can allow users to securely gain access to information spanning numerous data modalities, such as text, PDFs, images, and videos.

Enterprises can use NeMo Retriever to achieve more accurate results with less training, speeding time to market and supporting energy efficiency in the development of generative AI applications.

Reliable, Simple, Secure Deployment with NVIDIA AI Enterprise

Companies can deploy NeMo Retriever-powered applications to run during inference on NVIDIA-accelerated computing on virtually any data centre or cloud. NVIDIA AI Enterprise supports accelerated, high-performance inference with NVIDIA NeMo, NVIDIA Triton Inference Server™, NVIDIA TensorRT™, NVIDIA TensorRT-LLM, and other NVIDIA AI software.

To maximise inference performance, developers can run their models on NVIDIA GH200 Grace Hopper Superchips with TensorRT-LLM software.

Availability

Developers can sign up for early access to NVIDIA NeMo Retriever.

(0)(0)