NVIDIA AI Podcast How Visa Is Making Payments Safer and Smarter with AI - Ep. 256
45 snips
May 14, 2025 
Sarah Laszlo, Senior Director of Visa's Machine Learning Platform, shares her journey from psychology to shaping AI in finance. She explains how Visa uses AI to prevent fraud and enhance customer personalization. Laszlo discusses the shift from outdated code to Python, showcasing major cost savings through generative AI. She emphasizes the need for an unified AI strategy to combat fraud and stresses responsible AI practices, urging enterprises to foster collaboration between governance and technical teams.
AI Snips
Chapters
Transcript
Episode notes
AI Focuses on Fraud and Personalization
- AI's top priority at Visa is improving fraud prevention through even marginal gains.
- Visa is also advancing personalization and exploring agentic commerce for automated purchases.
Visa’s Unique Infrastructure Strategy
- Visa operates its own data centers rather than relying on cloud GPU resources due to privacy and security needs.
- This creates unique infrastructure challenges for deploying generative AI models at scale.
Maximize GPU Use with Virtualization
- Use virtual GPUs to split compute resources efficiently and improve cluster utilization.
- Giving users fractions of GPUs reduces waste and maintains performance for small jobs.