AI for Networking
From predictive internet systems to cognitive infrastructure
A clear line connects the Cisco-era work on predictive networking, endpoint analytics, wireless intelligence, and anomaly detection to the current AI4AI chapter.
Current work at NVIDIA
JP Vasseur currently focuses on advanced AI systems for autonomous data centers, AI factories, agent-native production, and the networking architecture required to operate large-scale AI infrastructure reliably.
Before 2024, his work centered on AI for networking: predictive internet systems, wireless analytics, endpoint intelligence, cognitive infrastructure, and the transition from reactive to predictive operations.
Ranked #79 on the worldwide prolific inventors list, with 769 patents and 626 patent families spanning networking, security, and ML/AI, including routing, QoS, MPLS, traffic engineering, optical networking, recovery, IoT, and voice/video.
Open patents page
Production AI systems, resilient infrastructure, and networking as a first-class part of large-scale AI environments.
Current NVIDIA Work
A curated view of published papers, videos, and forthcoming work across AI infrastructure, autonomous data centers, and AI factories.
Forthcoming
Upcoming papers centered on autonomous data centers, AI factories, and reliable production AI systems.
Updates
Stay informed when new papers, talks, and selected technical updates are published.
Subscribe here to receive new papers and selected updates. If you prefer a feed reader, the RSS feed is available below as well.
Prefer RSS? RSS feed
Research
Research remains visible here alongside current work, with the deeper archive collected on the dedicated research page.
AI for Networking
A clear line connects the Cisco-era work on predictive networking, endpoint analytics, wireless intelligence, and anomaly detection to the current AI4AI chapter.
Systems Thinking
Transport, telemetry, control, topology, and reliability remain core ingredients of AI platform design.
Trajectory
Protocol design, standards, and internet architecture continue to shape the way large-scale AI environments should be built and operated.
Cisco AI for Networking Journey
This is the earlier arc that shaped predictive networking, endpoint analytics, cognitive infrastructure, and the broader AI-for-networking viewpoint before the current NVIDIA chapter.
Neuroscience
This closes the homepage with one focused reference instead of a separate standalone chapter.
The brain is slow, unreliable, yet still the most complex, powerful, and energy-efficient device in the universe ...
Read white paper
Closing Note
JP Vasseur's personal website where he shares insights on networking, AI, machine learning, AI for networking, autonomous systems, security, IoT, and neuroscience.
LinkedIn: linkedin.com/in/jp-vasseur-phd Twitter/X: @jpvasseur
Disclaimer: any and all opinions and views expressed throughout the content of this website are JP Vasseur's own and shall not be deemed to reflect the views of any potential affiliates.