NVIDIA AI Foundry

Service · Custom development by Nvidia

Build custom generative AI models with your enterprise data and NVIDIA's expertise.

NVIDIA AI Foundry is a platform and service for building custom generative AI models with enterprise data and domain-specific knowledge. It provides access to NVIDIA's AI models, NeMo software for customization, and dedicated capacity on NVIDIA DGX Cloud, delivering custom models as NVIDIA NIM inference microservices for flexible deployment.

Category: Generative AI

Engagement model: Fixed project, Retainer, Time and materials, Subscription
Delivery: Remote, On-site, Hybrid
Team: NVIDIA AI experts, Specialized development teams
Customization Software: NVIDIA NeMo
Deployment Output: NVIDIA NIM inference microservice
Infrastructure: NVIDIA DGX Cloud

Build custom generative AI models with your enterprise data.
Leverage NVIDIA's foundation models and NeMo software for customization.
Deploy models as NVIDIA NIM inference microservices for flexible deployment.
Benefit from dedicated capacity on NVIDIA DGX Cloud with expert support.
Accelerate time to solution and protect data privacy.

Overview

NVIDIA AI Foundry is a comprehensive platform and service designed for enterprises to build custom generative AI models. It leverages proprietary enterprise data and domain-specific knowledge, functioning akin to a chip foundry by providing access to NVIDIA's AI models, the NeMo software for model customization, and dedicated capacity on NVIDIA DGX Cloud, all backed by NVIDIA AI experts. The service culminates in the delivery of NVIDIA NIM inference microservices, which include the custom model, optimized engines, and a standard API for flexible deployment. This enables organizations to develop their own AI models tailored to their specific needs, protecting data privacy and intellectual property while accelerating time to solution and achieving unprecedented performance.

Capabilities

Custom Model Development: Build custom generative AI models tailored for specific organizational, industrial, or cultural needs using enterprise data and domain expertise.
Foundation Models and Customization: Start with NVIDIA-created AI models like Nemotron and Edify, or popular open foundation models, and customize them using NVIDIA NeMo software.
Infrastructure and Expertise: Leverage dedicated capacity on NVIDIA DGX Cloud, built and backed by NVIDIA AI experts, for efficient model development.
Deployment Services: Output custom models as NVIDIA NIM inference microservices, complete with optimized engines and a standard API for flexible deployment anywhere.
Faster Time to Solution: Focus on leveraging organizational knowledge and AI insights without the burden of maintaining and tuning the AI development platform.
Unprecedented Performance: Build models efficiently on the latest NVIDIA architecture on DGX Cloud with preferred cloud service providers (CSPs), and deploy custom models with optimized engines packaged as NVIDIA NIM inference microservices.

What's included

Custom generative AI model development, Foundation model customization, NVIDIA DGX Cloud capacity, NVIDIA AI expert support, NVIDIA NIM inference microservice delivery

How it works

Augment and alter a pretrained foundation model using enterprise data and synthetically generated data. 2. Customize, evaluate, and apply guardrails to the model. 3. Output the customized model as an NVIDIA NIM inference microservice. 4. Use the NIM's standard API to build generative AI-powered app…

Use cases

Develop tailored generative AI solutions for specific industries and business needs.
Create custom AI models that leverage proprietary enterprise data and domain knowledge.
Accelerate the development and deployment of custom generative AI applications.
Enhance existing AI models with specialized data and functionalities.
Protect data privacy and intellectual property through custom model development.
Integrate custom AI models into enterprise applications via NVIDIA NIM microservices.
Improve AI model performance and accuracy for specific use cases.
Build AI agents capable of complex reasoning and action.

Specifications

Engagement model: Fixed project, Retainer, Time and materials, Subscription
Delivery: Remote, On-site, Hybrid
Team: NVIDIA AI experts, Specialized development teams
Industries served: Automotive, Financial Services, Healthcare, Manufacturing, Retail, Telecommunications, Media and Entertainment, Government, Energy, Consumer Goods, Industrial Manufacturing, Aerospace and Defense, Biotechnology, Pharmaceuticals, Higher Education, Research Institutions
Foundation Models Available: NVIDIA-created models (Nemotron, Edify), Popular open foundation models
Customization Software: NVIDIA NeMo
Deployment Output: NVIDIA NIM inference microservice
Infrastructure: NVIDIA DGX Cloud
Expertise: NVIDIA AI experts

How it works

Model Customization: Start with NVIDIA-created or popular open foundation models. Customize these models using NVIDIA NeMo software, incorporating enterprise data and domain expertise.
Evaluation and Guardrails: Evaluate the customized model and apply guardrails to ensure safety and alignment with business requirements.
NIM Microservice Packaging: The customized model is packaged as an NVIDIA NIM inference microservice, complete with optimized engines and a standard API.
Application Deployment and Improvement: Deploy the NIM microservice to build generative AI-powered applications. Feed knowledge gained from deployed applications back into the foundry to further improve custom models.

FAQ

What is NVIDIA AI Foundry?

NVIDIA AI Foundry is a platform and service for building custom generative AI models using your enterprise data and domain-specific knowledge. It provides access to NVIDIA's AI models, NeMo software for customization, and dedicated capacity on NVIDIA DGX Cloud, delivering custom models as NVIDIA NIM inference microservices.

What foundation models can I use with NVIDIA AI Foundry?

Enterprises can start with NVIDIA-created models like Nemotron and Edify, or popular open foundation models, and customize them using NVIDIA NeMo software. These models can then be deployed as NVIDIA NIM inference microservices.

What types of custom models can be built with NVIDIA AI Foundry?

NVIDIA AI Foundry enables the creation of custom models for various enterprise applications, including those requiring tailored language understanding, vision capabilities, speech processing, and healthcare-specific AI. Examples include customizing models for specific organizational, industrial, or cultural needs.

What infrastructure is used for model development?

NVIDIA AI Foundry provides dedicated capacity on NVIDIA DGX Cloud, which is built and backed by NVIDIA AI experts, ensuring efficient model development and access to state-of-the-art infrastructure.

What is the output of NVIDIA AI Foundry?

The output is an NVIDIA NIM inference microservice, which includes the custom model, optimized engines, and a standard API, allowing for flexible deployment across various environments.

How does NVIDIA AI Foundry help accelerate development?

NVIDIA AI Foundry helps accelerate the time to solution by allowing organizations to focus on leveraging their data and AI insights, rather than maintaining and tuning the AI development platform.