Dinis Guarda interviews Erik Pounds, Director of Product Marketing at NVIDIA in his YouTube Podcast series. In this episode, they talk about how NVIDIA is innovating in AI technology and making it accessible for enterprises, developers, and individuals. The podcast is powered by Businessabc.net and citiesabc.com.
Erik Pounds is the Director of Product Marketing at NVIDIA, focusing on AI and data science solutions. He is also the Senior Director of Enterprise AI at NVIDIA, promoting GPU-accelerated tools and technologies like the NVIDIA Merlin deep recommender system and the NVIDIA Riva conversational AI framework.
During the interview, Erik discusses with Dinis that NVIDIA is a leader in developing AI technologies, particularly in AI chips and models. NVIDIA’s GPUs (Graphics Processing Units) are widely used for training AI models due to their powerful processing capabilities. Their technologies support advancements in deep learning, machine learning, and data analytics.
Eric explains the vision of NVIDIA to Dinis:
“NVIDIA is focused on accelerated computing, especially in business, taking some computing tasks that are common and accelerating those to make them more efficient and drive cost down. Accelerated computing is a full stack challenge.
NVIDIA, as you know, is a semiconductor company or they bucket us that way. Well, the semiconductor, or the chip is just the bottom layer of the stack, it's a very important layer of the stack, but it's the bottom layer of the stack.
What sits above that are acceleration libraries to take advantage of that amazing computing power and then what sits above that are our domain specific software to apply that accelerated computing across various industry applications. Then, as you go further up the stack, you eventually end up getting to a specific application that takes advantage of everything that is below it and you do see NVIDIA climbing up the stack.
As we climb up the stack, we ensure that it's integrated with and it's compatible with all our partners and to make accelerated computing easier to use.”
Innovations at NVIDIA
Erik highlighted various innovations by NVIDIA. Speaking about the Enterprise AI division at NVIDIA, he says:
“You can think of NVIDIA AI Enterprise as a suite of software, like the operating system of AI that includes all the essential software you need to manage this kind of end to end pipeline of AI all the way from the start, where you curate your data to being able to run and scale these models in production. That's our only licensed software product. So, if you are a subscriber to NVIDIA AI Enterprise, you get all of NVIDIA's capabilities in the areas of AI and data science.”
Dinis and Erik discussed the need for augmenting proprietary data to train the foundation models, and ensure cross functionality.
“A lot of the foundation models are trained on essentially the knowledge of the world the internet and vast general purpose knowledge and skills sometimes they don't have the knowledge of of your community or your specific specific nation”, says Erik.
“Over the last two years, we've seen a lot of research into how we augment the knowledge of these AI models with external data. In a lot of cases, that's proprietary knowledge. That's why techniques like Retrieval Augmented Generation (or RAG) have been so hot in this space. Because these are the techniques now that allow you to take an amazing model. Even if you customise the model, you adapt that model to the core knowledge of your business, be an expert in the domains that you operate, and maybe have skills that you program.
When you augment that model with your proprietary knowledge, that's embedded in your data, which is your data, is a living breathing thing. Now, this model really becomes amazing and technically what happens is that the output and accuracy of the output of the model just goes up dramatically.”
He highlights that SAP has built on NVIDIA technology to bring solution for this challenge:
“SAP has built a co-pilot they call, Joule, which is available across all their kind of platforms - available to their developers, just to provide quick help and assistance in using your applications.”
Erik further comments:
“If you adopt NVIDIA software, you build your applications on top of our platform. You have our track record of forward and backwards compatibility, and this is very impressive which makes it easy to you know to adopt these solutions as well as run them across different types of systems.”
During the interview, Erik also explains the innovations in digital twin technology, and NVIDIA’s recently announced co-pilot Assistant:
“The human-like interfaces into these new applications is going to be very important because you can easily stand up a prototype chatbot by essentially just deploying a large language model, providing a text input interface in front of it and typing away.
But then, all the other amazing AI models that can essentially enhance that application are tremendous that are being taken advantage of by many early adopters today. For example, wouldn't it be great if you could just talk to it? Well, we have a suite of software, we call NVIDIA Riva.
It includes AI models for things like automatic speech recognition and text to speech going out. So, one can translate the coming speech and can return the speech going out, and even translate language, all in real time. So, now you can have more of a natural interface into the application.”
“The Omniverse connects the physical world to a virtual world using AI, all in real time. So, anything that you're doing that is physical, if you have a real time digital representation where you can be doing a lot of development and testing, you can just keep doing it in a virtual world and then deploy, iterate, test, and try new things out. Some things work, some things don't. So, instead of doing that in a physical world and taking an additional risk for your business or your application, you can do that in a virtual world.”
NVIDIA AI Foundry
NVIDIA has recently launched the NVIDIA AI Foundry with NVIDIA NIM™ inference microservices, designed to empower enterprises and nations to develop bespoke ‘supermodels’ designed to specific industry needs using the latest Llama 3.1 collection of openly available models.
Commenting on this, Erik tells Dinis:
“It's open, and can generate data that you can own and leverage in a commercial type application. One of the new interesting things about this model is that you can use it as it is and customise it.
One of the new applications is synthetic data generation. So, as you're building a custom model, you can take your proprietary data and use that to alter the model to have the knowledge of your domain, or the knowledge of your business, or you can even trade it, do certain skills in a lot of scenarios.
Nvidia AI Foundry gives you a service that you could use to take your proprietary data, bring that into The Foundry and customise these powerful models like Llama. We have evaluation tools as well so that you can evaluate the model across, whether it's industry benchmark or cross-benchmarks that are specific to your application. So, you ensure that you are building the best possible model and guardrail it so the model behaves in the way that you need it. We output models out of The Foundry in a stack that we call NVIDIA NIM that takes an optimised runtime and an easy to use API, wraps it all up into a container that you can deploy anywhere in the world, in the cloud, at the edge, in a device, on a workstation, and so on so forth.
So, now it makes building custom generative AI, its deployment, and use a whole lot easier.”
Concluding the interview, he highlights:
“We've gone past the days where everything is in the four walls of the business. Some of us still have to use VPNs, and things like this, to gain access to certain things, but a lot of these applications that we use today are behind a simple authentication mechanism. We can access that everywhere in the world, that was obviously a scary transition when those types of applications were rolling out, I'd say. It's the same for AI.
Start embracing these capabilities, then you'll learn what the things are that you need to do to best develop, scale and secure these new amazing powerful applications. They'll benefit your business tremendously.”
