NVIDIA Unveils New NIM Microservices for Governing AI at Scale
In an effort to help secure AI agents transforming the enterprise, NVIDIA is debuting three new NVIDIA NIM microservices for AI guardrails, joining the company’s existing NeMo Guardrails collection of software tools. Serving to increase the accuracy, security, and control of proprietary AI implementations, NVIDIA continues to promote trustworthy, governed AI amid constant innovation.
AI agents are a forerunner in popular AI implementations, potentially able to transform productivity for knowledge workers everywhere. As these agents move past their development and experimentation phases into deployment and production, ensuring that these agents are performant, fast, and secure is paramount.
NeMo Guardrails is a scalable platform for defining, orchestrating, and enforcing AI guardrails in both agentic use cases and generative AI (GenAI). The latest AI guardrails “help maintain the credibility and the reliability of AI operations by enforcing specifications for AI models, agents, and actual systems,” explained Kari Briski, vice president, generative AI software, product management at NVIDIA. “In other words, it helps keep AI agents on track.”
This is accomplished by simplifying the processes of applying multiple specialized, light-weight rails—or AI software policies—onto AI and large language models (LLMs), which are customizable based on an enterprise’s unique use case. The configurability of these microservices allows developers to apply niche rules to their complex AI workflows, filling the gaps left by more general global policies.
These small language models are designed to offer lower latency and run efficiently, even in distributed or resource-constrained environments—which is particularly useful for industries such as healthcare, automotive, and manufacturing, and in locations like hospitals or warehouses, according to NVIDIA.
NVIDIA’s new microservices include:
“These three new NIM microservices offer needed layers of security and control that our customers are asking for so that they can deploy these agentic AIs into their applications,” said Briski. “[Because] they're GPU accelerated…[as] NIMs…they perform with low latency and can be securely deployed anywhere as part of NVIDIA AI enterprise.”
With NeMo Guardrails available to the open source community, some enterprises are already realizing the benefits of NVIDIA’s new microservices.
“Technologies like NeMo Guardrails are essential for safeguarding generative AI applications, helping make sure they operate securely and ethically,” said Anthony Goonetilleke, group president of technology and head of strategy at Amdocs, who is implementing NeMo Guardrails. “By integrating NVIDIA NeMo Guardrails into our amAIz platform, we are enhancing the platform’s ‘Trusted AI’ capabilities to deliver agentic experiences that are safe, reliable, and scalable. This empowers service providers to deploy AI solutions safely and with confidence, setting new standards for AI innovation and operational excellence.”
“Cerence AI relies on high-performing, secure solutions from NVIDIA to power our in-car assistant technologies,” said Nils Schanz, executive vice president of product and technology at Cerence AI, another user of NeMo Guardrails. “Using NeMo Guardrails helps us deliver trusted, context-aware solutions to our automaker customers and provide sensible, mindful, and hallucination-free responses. In addition, NeMo Guardrails is customizable for our automaker customers and helps us filter harmful or unpleasant requests, securing our CaLLM family of language models from unintended or inappropriate content delivery to end users.”
To learn more about NVIDIA’s NIM microservices for AI guardrails, please visit https://www.nvidia.com/en-us/.