NVIDIA Unveils AI Vision at Money20/20 USA
In an exclusive interview with FinTech Magazine at Money20/20, Pahal Patangia, Head of Global Industry Business Development, Payments at NVIDIA, details the company's strategic approach to revolutionising financial services through its full-stack accelerated computing platform.
The Evolution of AI in Finance
The transformation of AI in financial services has its roots in a pivotal moment a decade ago. “If you look back 10 to 15 years ago, it was in 2012 when AlexNet was released. That's when the world woke up from the AI winter,” says Pahal.
“Fast forward to 2022 when ChatGPT emerged, and that was the watershed moment for the entire community to see what this new wave is all about.”
This evolution marks a significant shift in how financial institutions handle data. “We're now thinking of datasets moving from terabyte to petabyte scale, and soon will be reaching exabyte scale,” Pahal notes.
The implications of this growth are far-reaching. “Every segment of financial services – from capital markets to banking and payments – is excited about this wave and bringing new use cases to leverage this technology,” he continues.
“We've released a fraud detection workflow that enterprises can use as a framework to build upon, which is available and deployable across any platform"
Building a Three-Layer Computing Platform
NVIDIA's journey in accelerated computing began in 2006 with the launch of CUDA, a framework designed to parallelise applications and workloads that traditionally required significant processing time.
The company has since developed what Pahal describes as a ‘three-layer cake’ approach to accelerated computing.
The foundation layer comprises hardware – chips and systems – connected through NVLink and InfiniBand technologies, significantly enhanced through the Mellanox acquisition in 2020.
This integration creates a robust system-level computing foundation that serves as the bedrock for AI applications.
The middle layer features NVIDIA AI Enterprise, which Pahal describes as “the operating system for all things AI”.
This suite of software libraries is engineered to extract maximum acceleration from the hardware layer, ensuring optimal performance for AI workloads.
The top tier consists of application frameworks, providing enterprises with the tools to develop and retain IP for their specific use cases while maintaining control over their technological assets.
Transforming Financial Data Analysis
The financial services industry stands at a crucial juncture in its data analysis capabilities. “Traditionally, this industry has heavily relied on structured data for insights,” Pahal explains.
“But now, all the data stored in SharePoint – product guidelines, customer call transcripts – has opened up new possibilities for serving customers better and build curated products for them.”
This convergence of structured and unstructured data is already yielding results in major financial institutions.
“We had BlackRock present at NVIDIA GTC, where they demonstrated leveraging unstructured data through news articles, using knowledge graphs and retrieval-augmented generation to produce insights using NVIDIA’s accelerated computing platform,” adds Pahal.
To facilitate broader adoption of these capabilities, NVIDIA has developed practical solutions for the industry. The company has introduced an end-to-end fraud detection workflow, which Pahal describes as a framework that enterprises can build upon.
Using a combination of advanced algorithms, one of which tapped into the NVIDIA AI platform, American Express enhanced model accuracy for fraud detection. PayPal optimised its data processing pipeline using Spark-RAPIDS which comes as a part of NVIDIA AI Enterprise, a key component in the fraud detection workflow, achieving a 70% cost reduction with NVIDIA accelerated computing.
“We've released a fraud detection workflow that enterprises can use as a framework to build upon, which is available and deployable across any platform. These enablements we're seeing are just the beginning of what's possible when you tie use cases to the platform,” concludes Pahal.
Learn more about the NVIDIA AI workflow for fraud detection here.
Explore the latest edition of FinTech Magazine and be part of the conversation at our global conference series, FinTech LIVE.
Discover all our upcoming events and secure your tickets today.
FinTech Magazine is a BizClik brand
