INTERNET INFRASTRUCTURE · EDGE AI
Tucker
Preston.
Global networks.
Distributed intelligence.
I’ve spent my adult career making the Internet faster at the edge. My work brings together global connectivity, peering and traffic engineering—first for CDN and cloud services, and in recent years for distributed AI inference.
At Gcore, I lead global connectivity, peering and traffic engineering for CDN, cloud and distributed AI infrastructure.
I’m passionate about Internet measurement: understanding how networks behave at scale. My research and engineering interests span AI inference optimization, routing and global network operations.
Read my field notes ↗︎About meENGINEERING FOCUS
Backbone architecture & peering
Leading backbone design and procurement, alongside global connectivity and interconnection.
Traffic engineering & steering
BGP, Anycast and routing decisions that connect performance, capacity and resilience.
AI & distributed inference
Connectivity for GPU infrastructure, distributed model serving and AI services across regions and deployment environments.
FLAGSHIP PROJECT / GCORE
A global backbone.
Built to reach further.
I lead Gcore’s global private backbone initiative. I designed the network and procured its connectivity, connecting architecture decisions with the infrastructure needed to deliver them.
My role in the project ↗︎CURRENT FOOTPRINT
Asia-Pacific
Europe
North America
EXPANDING SOON
Latin America
Africa
Australia
100G / 400G Dedicated wavelengths
Read my article on Gcore ↗︎FIELD NOTES / ORIGINAL WRITING
All notes ↗︎Ideas from the network.
Measuring Isaias.
Through
the network.
Measuring Hurricane Isaias through Internet reachability. Coastal losses and uneven recovery, seen from two independent measurement points.
Explore the measurements
THE GULF COAST / OBSERVED REACHABILITYWorst complete-hour decline in replies per probe
AI INFERENCE & ROUTING
The network path behind an AI response
An inference request crosses several systems before the first token appears. Knowing where it waits tells you what to fix.
Read note ↗︎02EDGE INFRASTRUCTURE
Edge inference sites: interconnection before distance
A nearby building can still have a long network path to its users. Site diligence needs to establish who you can reach, how you reach them and what happens when a connection fails.
Read note ↗︎SELECTED WRITING & CONTRIBUTIONS
View all selected work ↗︎Why is IPv6 adoption slow?
The New Stack · Gcore-sponsored technical article
↗︎02 / EMPLOYER INTERVIEWWhere CDN infrastructure is heading
Gcore · industry interview, October 2024
↗︎03 / RESEARCH & ENGINEERINGAI inference, routing & Internet measurement
Engineering interests, related reading and research contributions
↗︎UP NEXT / OCTOBER 19–21, 2026Meet me at NANOG 98
Miami, Florida · Peering, networking and AI infrastructure
↗︎