NVIDIA NeMo
An agent-first, open suite of libraries with skills for accelerating AI agent specialization, optimization, and governance.
Overview
What Is NVIDIA NeMo?
NVIDIA NeMo™ is an agent-first, open suite of libraries with skills for accelerating AI agent specialization, optimization, and governance.
NeMo integrates with existing AI tools and agent frameworks to optimize specialized agents across any cloud, on-premises, or hybrid environment.
Features
Tools and Skills for Accelerating Specialized AI Agent Optimization
The AI agent lifecycle is an end-to-end process for developing and continuously improving AI agents in production applications. NeMo integrates with existing AI tools and agent frameworks to optimize specialized agents across their lifecycles.
| Build | |
|---|---|
| Prepare AI-ready data Process existing multimodal datasets into high-quality, AI-ready formats for development pipelines, and generate synthetic data to close critical data gaps. |
NeMo Curator Clean, filter, and prepare safe multimodal data for agentic use cases and model training. Create domain‑specific synthetic datasets from scratch for building and evaluating specialized agents. Perform context-aware data anonymization to protect PII while preserving insights. Generate safe, synthetic versions of your sensitive datasets with no one-to-one mapping to original records.
|
| Select the right modell Pick or build models suited to the use case: selecting from Nemotron™ or other open models, other open or proprietary options, or training from scratch. Validate with evaluation runs, and fine-tune as needed. |
NVIDIA Nemotron State-of-the-art open NVIDIA models for reasoning, RAG, speech, vision, and safety. Benchmark and test models and agents using academic, LLM-as-a-judge, and custom evaluations.
|
| Deploy | |
| Deploy your agent with maximum performance Optimize your agent for production with high-throughput, low-latency inference, ensuring it can scale to meet enterprise demands and deliver fast, reliable responses. |
NVIDIA NIM Run AI models in optimized containers, exposed as OpenAI-compatible APIs. " data-col-index="1" data-row-index="5" class=" even-column even-row">
|
| Stay grounded in data and apply guardrails Use retrieval-augmented generation (RAG) to anchor agent responses in trusted knowledge while applying safety, compliance, and content moderation guardrails. |
NeMo Guardrails Enhance safety, compliance, and control across AI interactions. " data-col-index="1" data-row-index="6" class=" even-column odd-row">
|
| Validate agent and model safety before launch Identify and remediate safety and security vulnerabilities in models and agents before they reach production. |
NeMo Auditor Identify safety and security vulnerabilities. " data-col-index="1" data-row-index="6" class=" even-column odd-row">
|
| Optimize | |
| Profile and Optimize Your Agent Track the agent's real-world interactions with users and other systems. Systematically evaluate its performance and accuracy, finding opportunities to continuously improve. |
NeMo Relay Connect and observe black-box and general-purpose agent harnesses into the NeMo platform. Microservice to fine-tune and align models with domain data. Collection of all open-source NeMo libraries for training and aligning LLMs and multimodal models efficiently at scale.
|
| Continuously improve with data flywheels Use the feedback and data gathered from monitoring to create a data-driven flywheel, iteratively retraining the agent to continuously optimize and stay effective over time. |
NeMo RL Post-train and align models at scale with advanced reinforcement learning (RL) techniques. Simulated RL training environments to generate high-quality agentic RL training data/rollouts. Benchmark and test models and agents using academic, LLM-as-a-judge, and custom evaluations. Plug-in to fine-tune and align models with domain data.
|
Use Cases
How NeMo Is Being Used
See how NVIDIA NeMo supports industry use cases and jump-starts your AI development.
-
AI Agents
-
SDG for Agentic AI
-
AI Assistant
-
Enterprise Search
-
Content Generation
-
Humanoid Robot
AI Agents
AI agents are transforming customer service across sectors, helping companies enhance customer conversations, achieve high resolution rates, and improve human representative productivity. AI agents can handle predictive tasks, reason and problem-solve, be trained to understand industry-specific terms, and pull relevant information from an organization’s knowledge bases, wherever that data resides.
Synthetic Data Generation for Agentic AI
Specialized agentic systems need massive, high-quality datasets that are slow and expensive to collect from real-world sources. Synthetic data created through simulations or generative AI models can eliminate this bottleneck by creating unlimited training scenarios without privacy restrictions or quality issues. This enables faster development of reasoning LLMs, multi-step decision-makers, and multimodal AI assistants.
AI Assistant
Businesses are deploying AI assistants to efficiently address the queries of millions of customers and employees around the clock. Powered by customized NVIDIA NIM™ microservices for LLMs, RAG, and speech and translation AI, these AI teammates deliver immediate and accurate spoken responses, even in the presence of background noise, poor sound quality, and diverse dialects and accents.
Enterprise Search
Enterprises generate trillions of documents annually—including PDFs, reports, presentations, —each containing text, images, charts, and tables—spread across disconnected systems. AI-powered enterprise search transforms this scattered data into a unified knowledge base, enabling employees to instantly surface insights using natural language and driving faster decisions at lower cost.
Content Generation
Generative AI makes it possible to generate highly relevant, bespoke, and accurate content grounded in the domain expertise and proprietary IP of your enterprise.
Humanoid Robot
Humanoid robots are built to adapt quickly to existing human-centric urban and industrial work spaces, tackling tedious, repetitive, or physically demanding tasks. Their versatility has them in such varied locations as factory floors to healthcare facilities, where these robots are assisting humans and helping alleviate labor shortages with automation.
Apptronik
Benefits
Explore the Benefits of NVIDIA NeMo for Agentic AI
Agent-First Development With Skills
Manage the full agent lifecycle from data curation and post-training to evaluation, guardrails, observability, and continuous optimization using an agent-friendly suite of agent skills.
Accelerate at Scale
Deploy and scale data flywheels using enterprise data, with GPU-accelerated training, inference, multi-node scaling, and cost-efficient optimization for high-throughput agent workloads.
Increased ROI
Build, customize, and deploy specialized agentic systems faster—shortening time to production and maximizing return on AI investments.
Secure and Production-Ready
Safeguard sensitive data, enforce policy and prompt guardrails, validate models, and continuously detect vulnerabilities. Deploy securely with enterprise-grade support and stability across cloud, data center, and edge with NVIDIA AI Enterprise.
Customer Stories
How Industry Leaders Are Driving Innovation With NeMo
Adopters
Leading Adopters Across All Industries
-
Customers
-
Partners
Resources
The Latest in NVIDIA NeMo Resources
-
Blogs
-
Sessions
-
Training
-
Videos
Next Steps
Ready to Get Started?
Use the right tools and technologies to take your agentic AI applications from development to production.
For Developers
Explore everything you need to start developing with NVIDIA NeMo, including the latest documentation, tutorials, technical blogs, and more.
Get in Touch
Talk to an NVIDIA product specialist about moving from pilot to production with the assurance of security, API stability, and support that comes with NVIDIA AI Enterprise.
Shell
Shell Trains Custom AI Chatbot With NVIDIA NeMo to Uplevel Operations
Shell, a global leader in the energy industry, has leveraged NVIDIA NeMo to empower its journey toward developing a custom AI chatbot for chemical domain expertise. This innovative solution has the potential to significantly enhance employee productivity by streamlining search processes, improving decision-making, and supporting research and development in production environments.
AI Sweden
Accelerate Industry Applications With LLMs
AI Sweden facilitated regional language model applications by providing easy access to a powerful 100 billion-parameter model. They digitized historical records to develop language models for commercial use.
Amazon
How Amazon and NVIDIA Help Sellers Create Better Product Listings With AI
Amazon doubles inference speeds for new AI capabilities using NVIDIA TensorRT-LLM and GPUs to help sellers optimize product listings faster.
Amdocs
NVIDIA and Amdocs Bring Custom Generative AI to Global Telco Industry
Amdocs plans to build custom LLMs for $1.7 trillion global telecommunications industry using NVIDIA AI foundry service on Microsoft Azure.
AT&T
AT&T Drives Customer Care AI Agents’ Accuracy, Efficiency, and Performance With NVIDIA NeMo
AT&T, one of the world’s largest telecommunications companies, is reimagining customer care through the power of AI. Facing challenges like model drift, rising computational demands, and the need for real-time data access, AT&T turned to NVIDIA NeMo™ microservices to build a feedback-driven AI platform that continuously improves performance while optimizing cost, speed, and compliance.
AWS
NVIDIA Powers Training for Some of the Largest Amazon Titan Foundation Models
Amazon leveraged the NVIDIA NeMo framework, GPUs, and AWS EFAs to train its next-generation LLM, giving some of the largest Amazon Titan foundation models customers a faster, more accessible solution for generative AI.
Accenture
Accelerate Generative AI Adoption for Enterprises
ServiceNow, NVIDIA, and Accenture announced the launch of AI Lighthouse, a first-of-its-kind program designed to fast-track the development and adoption of enterprise generative AI capabilities.
Azure
Harnessing the Power of NVIDIA AI Enterprise on Azure Machine Learning
Get access to a complete ecosystem of tools, libraries, frameworks, and support services tailored for enterprise environments on Microsoft Azure.
Bria
Bria Builds Responsible Generative AI for Enterprises Using NVIDIA NeMo, Picasso
Bria, a startup based in Tel Aviv, is helping businesses who are seeking responsible ways to integrate visual generative AI technology into their enterprise products with a generative AI service that emphasizes model transparency alongside fair attribution and copyright protections.
Cohesity
Unlock Your Data Superpower: NVIDIA Microservices Unleash Enterprise-Grade Secure Generative AI for Cohesity
With NVIDIA NIM and optimized models, Cohesity DataProtect customers can add generative AI intelligence to data backups and archives. This allows Cohesity and NVIDIA to bring the power of generative AI to all Cohesity DataProtect customers. Leveraging the power of NIM and NVIDIA optimized models, Cohesity DataProtect customers obtain the power of data-driven insights from their data backups and archives, unleashing new levels of efficiency, innovation, and growth.
CrowdStrike
Shaping the Future of AI in the Cybersecurity Domain
CrowdStrike and NVIDIA are leveraging accelerated computing and generative AI to provide customers with an innovative range of AI-powered solutions tailored to efficiently address security threats.
Dell
Dell Validated Design for Generative AI With NVIDIA
Dell Technologies and NVIDIA announced an initiative to make it easier for businesses to build and use generative AI models on premises quickly and securely.
Deloitte
Unlock the Value of Generative AI Across Enterprise Software Platforms
Deloitte will use NVIDIA AI technology and expertise to build high-performing generative AI solutions for enterprise software platforms to help unlock significant business value.
Domino Data Lab
Domino Offers Production-Ready Generative AI Powered by NVIDIA
With NVIDIA NeMo, data scientists can fine-tune LLMs in Domino’s platform for domain-specific use cases based on proprietary data and IP—without needing to start from scratch.
Dropbox
Dropbox and NVIDIA to Bring Personalized Generative AI to Millions of Customers
Dropbox plans to leverage NVIDIA’s AI foundry to build custom models and improve AI-powered knowledge work with Dropbox Dash universal search tool and Dropbox AI.
Google Cloud
AI Titans Collaborate to Create Generative AI Magic
At its Next conference, Google Cloud announced the availability of its A3 instances powered by NVIDIA H100 Tensor Core GPUs. Engineering teams from both companies have collaborated to bring NVIDIA NeMo to the A3 instances for faster training and inference.
HuggingFace
Leading AI Community to Accelerate Data Curation Pipeline
Hugging Face, the leading open platform for AI builders, is collaborating with NVIDIA to integrate NeMo Curator and accelerate DataTrove, their data filtering and deduplication library. “We are excited about the GPU acceleration capabilities of NeMo Curator and can’t wait to see them contributed to DataTrove!” says Jeff Boudier, Product Director at Hugging Face.
KT
Creating New Customer Experiences With LLMs
South Korea’s leading mobile operator builds billion-parameter LLMs trained with the NVIDIA DGX SuperPOD platform and NeMo framework to power smart speakers and customer call centers.
Lenovo
New Reference Architecture for Generative AI Based on LLMs
Solution to expedite innovation by empowering global partners and customers to develop, train, and deploy AI at scale across industry verticals with utmost safety and efficiency.
Quantiphi
Enabling Enterprises to Fast-Track Their AI-Driven Journeys
Quantiphi specializes in training and fine-tuning foundation models using the NVIDIA NeMo framework, as well as optimizing deployments at scale with the NVIDIA AI Enterprise software platform, while adhering to responsible AI principles.
SAP
SAP and NVIDIA Accelerate Generative AI Adoption Across Enterprise Applications Powering Global Industries
Customers can harness their business data in cloud solutions from SAP using customized LLMs deployed with NVIDIA AI foundry services and NVIDIA NIM Microservices.
ServiceNow
Building Generative AI Across Enterprise IT
ServiceNow develops custom LLMs on its ServiceNow platform to enable intelligent workflow automation and boost productivity across enterprise IT processes.
Perplexity
Enhance Model Performance for AI-Powered Search Engines
Using NVIDIA NeMo, Perplexity aims to quickly customize frontier models to improve the accuracy and quality of search results and optimize them for lower latency and high throughput for a better user experience.
VMware
VMware and NVIDIA Unlock Generative AI for Enterprises
VMware Private AI Foundation with NVIDIA will enable enterprises to customize models and run generative AI applications, including intelligent chatbots, assistants, search, and summarization.
Weight & Biases
Debug, Optimize, and Monitor LLM Pipelines
Weights & Biases helps teams working on generative AI use cases or with LLMs track and visualize all prompt-engineering experiments—helping users debug and optimize LLM pipelines—as well as provides monitoring and observability capabilities for LLMs.
Writer
Startup Pens Generative AI Success Story With NVIDIA NeMo
Using NVIDIA NeMo, Writer is building LLMs that are helping hundreds of companies create custom content for enterprise use cases across marketing, training, support, and more.
Arize
Arize Powers Self-Improving AI Data Flywheels
Arize’s LLM engineering and observability platform integrates NVIDIA NeMo microservices to power AI data flywheels, enabling continuous model refinement through real-world feedback. With NeMo Customizer, Evaluator, and Guardrails, Arize ensures agentic systems are performant, safe, and aligned with evolving enterprise needs. This collaboration supports the development of adaptive AI that learns and evolves over time.
DataRobot
Enterprise-Ready, Trustworthy AI Agents With NeMo on DataRobot
With NVIDIA NeMo embedded into the DataRobot Enterprise AI Suite, enterprises can ensure agentic systems are safe, compliant, and grounded in enterprise-specific data. This integration facilitates the development of AI agents that deliver accurate, context-aware responses while adhering to organizational standards.
DataStax
DataStax and NVIDIA Build Data and AI Platform
Over the past year, DataStax has partnered with NVIDIA to adopt NVIDIA NeMo microservices to enhance generative AI, retrieval-augmented generation, and hybrid search across its database and AI offerings. The results have been impressive: 19x better performance in throughput, a significant reduction in costs, and improved latency.
Galileo
Galileo and NVIDIA NeMo: De-Risking Agentic AI in Production
Galileo integrates NVIDIA NeMo microservices to build AI data flywheels that strengthen agent performance, reliability, and trust. NeMo adds complementary capabilities to the Galileo platform—enabling continuous domain-specific fine-tuning via NeMo Customizer, advanced evaluation with NeMo Evaluator, and safeguarding user interactions with NeMo Guardrails to empower AI teams to build, evaluate, and monitor agentic AI systems that learn and improve continuously in real-world environments.
