MONEY20/20: SQream Showcases GPU Breakthrough

Share
MONEY20/20: SQream Showcases GPU Breakthrough
SQream supports NVIDIA & Oracle to improve data processing speeds, reducing analysis time & changing how enterprises handle structured data workflows

In an exclusive interview at Money20/20 USA 2024, SQream's leadership reveals how its partnership with tech giants NVIDIA and Oracle is reshaping enterprise data processing capabilities and accelerating time to insight across the financial services sector.

Youtube Placeholder

Breaking Through the CPU Ceiling

The selection of SQream by NVIDIA and Oracle to showcase its GPU-powered data acceleration technology at Money20/20 USA 2024 underscores a pivotal shift in enterprise computing. 

The company's innovative approach addresses a critical market need: processing enormous volumes of structured data at unprecedented speeds, particularly in financial services where traditional CPU-based infrastructure has reached its limits.

“By adding SQream into the pipeline, we shattered the bottlenecks through GPU acceleration and shrunk that pipeline by over 80%”

Deborah Leff, SVP of Data Science and Machine Learning at SQream

“When you have an extended data pipeline and absolutely need to get to that insight faster, SQream can provide an order of magnitude improvement,” says Deborah Leff, SVP of Data Science and Machine Learning at SQream. 

This claim is substantiated by a recent implementation at a major bank serving 50-60 million daily active online banking customers, where the institution struggled to update its fraud models daily due to CPU infrastructure limitations.

Revolutionising Financial Services Processing

The implementation of SQream's technology has demonstrated transformative results in real-world applications. 

“By adding SQream into the pipeline, we shattered the bottlenecks through accelerated computing and shrunk that pipeline by over 80%,” Deborah explains. “What previously took 37 hours to complete is now done in seven hours. Now, if you kick off an update every night at midnight, by the time the next workday begins, all your models will be updated.”

This breakthrough stems from SQream's 14-year partnership with NVIDIA, which has yielded a software solution capable of running on virtually all NVIDIA hardware. 

The technology enables data processing with accelerated computing across various enterprise environments, whether on-premises or in the cloud, marking a significant advancement in data processing capabilities.

SQream has leveraged NVIDIA and Oracle’s capabilities to pioneer an innovative approach to address critical market needs in financial services.

This includes improving the accuracy of fraud detection systems to better protect organisations’ customers and merchants, removing latency related to processing month-end closes (which improves business insights for executives), and enhancing customer service by providing dramatically faster time to relevant insights. 

Deborah Leff, SVP of Data Science and Machine Learning at SQream

Strategic Partnerships Driving Innovation

The collaboration with Oracle adds another dimension to SQream's offering. Oracle's AI-centric cloud infrastructure provides an ideal platform for SQream's technology, creating what Deborah describes as “the most efficient, effective way to break through bottlenecks and create tremendous value from structured data”.

Deborah illustrates the paradigm shift by analogy to transportation evolution: “Every service level agreement has been built on what a CPU can do because that's all we had. It's like before planes were invented, everything was a car. If the journey took two hours or seven days, that's all you had. 

“Then along comes an aeroplane, and now instead of seven days in a car, you can get there in just a few hours. It’s like driving from New York to Money20/20 in Las Vegas, you’ll get there in the end, but booking a seat on a plane will get you there much faster.

"Catching a flight for such a journey is the new norm, and the same goes for leveraging GPU resources for structured data processing. Once you experience the speed and efficiency, there's no going back, especially when time is of the essence."

**************

Make sure you check out the latest edition of FinTech Magazine and also sign up to our global conference series – FinTech LIVE 2024

**************

FinTech Magazine is a BizClik brand. 

Share

Featured Articles

The FinTech Year in Stories: April

We look at the articles that made the news in fintech in 2024. Today, it is April…

The FinTech Year in Stories: March

We look at the articles that made the news in fintech in 2024. Today, it is March…

The FinTech Year in Stories: February

We look at the articles that made the news in fintech in 2024. Today, it is February…

The FinTech Year in Stories: January

Financial Services (FinServ)

FinTech Predictions for 2025 - Pt. 2

Financial Services (FinServ)

Fintech Predictions for 2025 – Pt.1

Digital Payments