Nvidia with Kevin Levitt | E199

Powering the future of Artificial Intelligence.

In today’s episode, Jason is going to talk to Kevin Levitt; he is the Director of Industry and Business Development for Financial Services for Nvidia. The company is one of the driving forces behind the technology powering artificial intelligence today.

Episode Highlights:

  • 00.38: Nvidia is a company that was started about 30 years ago almost, and they have really pioneered the use of graphics processing units. GPU’s initially for rendering images and videos and graphics and everything that you think of a great portion of our Business today and our heritage which is in the graphics industry, video games, etc. fast forward to today, and not only do we power some of the most amazing visual platforms and capabilities imaginable. 

  • 01.11: In Nvidia, we empower all the artificial intelligence capabilities that are resident within any industry, whether that is automotive healthcare and certainly financial services, says Kevin. 

  • 03.35: Jason says there is no one specific type of artificial intelligence in computing.

  • 03.42: Kevin gives the listener a brief introduction into the differences between AI deep learning and machine learning.

  • 04.24: The deep learning algorithm really identifies whether a prediction is accurate or not through its own neural network learning and improving on its own. You can think of deep learning as software writing software, says Kevin.

  • 05.10: Jason says for a long-term time, artificial intelligence has been an abstract concept. 

  • 06.37: After product discovery, it is all about going in applying for the for the product, says Kevin.

  • 07.40: Jason talks about the four seas collateral character capacity and credit. That is a very narrow framing of a problem. When you know what these guys have done clearly from what you are describing, they have taken in countless variables to create models.

  • 08.12: What Jason has seen in academic studies is more accurate FICO scores in terms of calculating the probability of default.

  • 08.43 Kevin says there are myriad applications, and underwriting these cases doesn’t apply, just to credit. It applies to insurance where companies are looking at how we drive versus the data coming to them from third parties to use telematics to see that.

  • 09.58 Jason asks, “What is the natural type of function for artificial intelligence disturbing the market today like what is the commonality around the things is replacing?

  • 10.20: Kevin says in the example of Siri virtual assistant or chatbot. In the context of financial services that are helping us to transfer a balance or to understand what our balance is, or pay a bill, it goes from there to assist the call center agent where we have a more complex problem. With the call center and the agent, the AI is actually complimenting human assistants with information.

  • 11.20: A lot of what we see with financial services is grounded in natural language capabilities, extending beyond a consumer finance, says Kevin.

  • 11.52 Kevin: Natural language processing models are all capable of either trading algorithmically or again aiding the human trader and identifying signal that is important to their investment decisions. 

  • 12.53 Jason: There are large quantities of data that are basically analyzed in a systematic way, and it is really just a ton of heavy lifting beyond the capacity of the human mind, that sort of thing is far probably are easily handled by an AI than any human being ever could.

  • 13.50: Kevin highlights one of the areas where they are seeing a lot of investment from their customers. It is in identifying fraudulent transactions, and the AI is actually proven to be really good here.

  • 14.15: AI is not good at context switching. For example, you could train them to drive a vehicle and drive a car, but then if you put them in a different type of. Mode of transportation, such as a train. They’re not going to be able to context switch immediately and do that. Humans can configure out well. We would still need to find a way to accelerate, says Kevin.

  • 16.50: Consultancy’s are full of huge teams. Rather than analyzing spreadsheets, they’re analyzing the output and the data to make better-informed decisions and consult the companies hiring them more effectively.

  • 18.30: Large banks are trying to figure out how to build an enterprise AI capability, AI infrastructure to support the migration from a handful of AI-enabled applications up to 100s.

  • 19.14: One of the more challenging areas in building out an AI capability within the large banks, even within the fintech’s, is how do you recruit and retain all this highly desirable talent in terms of the data scientists themselves, says Kevin.

  • 20.51: As per Kevin, it is not just about the talent, it is not just about the infrastructure, it is about identifying the use cases that where AI will most benefit both the bank and other financial institutions, as well as their customers. 

  • 22.10: Jason inquires, “What is the kind of cool use cases you see being drummed up and coming forward going in the future?” 

  • 24.40: You can’t expect people to do everything you are told. The more the system can do it for them, the better off you are and will call it a construction or concept of benevolent nudging that can really impact people’s lives, says Jason.

  • 26.30: Kevin talks about the four primary players in terms of big retail, big tech, fintech and big banks, are going to be the primary competitors and if one of them is using AI to deliver a virtual assistant or chatbot and the other one is still using some form of rules-based kind of chat experience, AI one is going to win.

  • 28.40 Jason: The technology companies choosing to come out and this is going to make everybody sharper, and everybody really focused on their value proposition and really try to eliminate friction.

  • 30.49: Kevin says there are all sorts of embedded finance capabilities that will not require machine learning or deep learning to exist. But from a customer experience standpoint, they will make a huge impact in kind of our day-to-day life.

  • 32.05: AI is so paramount to this industry whether it’s for the fintech’s, the insurer tax, the big banks, and for any company that is going to be investing in AI and migrating into deep learning capabilities you need an accelerating computing platform.

  • 35.15 Kevin remarks it is super exciting every day to wake up to know that we can address so many problems and really work to solve the most challenging problems. 

  • 35.32: NVIDIA is all about innovation and stretching, kind of the boundaries of where people thought. Computing power could go and certainly where artificial intelligence could be of benefit.

 

3 Key Points:

  1. For the past 15 plus years, Kevin has been at the intersection of data technology and financial services.

  2. The technology can enable a better customer experience across many dimensions when artificial intelligence and deep learning models that leverage natural language processing are utilized.

  3. There are lots of opportunities to continually improve how AI is leveraged within any industry, including within the context of financial services.

Tweetable Quotes:

  • “You can think of artificial intelligence, or AI is kind of the Super umbrella if you will, and underneath that falls a category of artificial intelligence which is machine learning.” – Kevin

  • “You talked about natural language processing, natural language understanding, and we’re all using language, primarily English, but many other languages in this country.” - Kevin 

  • “It is not about job loss it is about job improvement, which is freeing us to do the higher-order capabilities.” – Kevin

  • “There are some of the smartest people in the world that are working on financial services, and they see the power and the opportunity associated with AI.” - Kevin

Resources Mentioned: