Operating history / distributed systems

Infrastructure for billions of users and millions of requests per second

A carefully attributed account of Francisco Perez Romero's work across global routing, telemetry, CDN, capacity planning, reliability, performance, and AI infrastructure.

Book an assessment-fit call ↗

By: Francisco Perez Romero · 2026-08-07

99.999%

reliability context

Millions/sec

reverse-CDN request scale

4→15+

team growth

18+

years of experience

Decision guide

Evidence before prescription.

Meta: global traffic infrastructure

Francisco built and led networking infrastructure used across Facebook properties. The work included telemetry pipelines, traffic simulation, capacity planning, real-time routing, performance optimization, and technical leadership as the team grew from 4 to 15+ engineers.

Telemetry powering routing decisions.
Simulation for proactive capacity planning.
Architecture and operations at global traffic scale.
Management while remaining hands-on with long-term technical work.

Microsoft: CDN and the foundations of Azure networking

At Microsoft, Francisco worked on reverse-CDN infrastructure handling millions of requests per second, telemetry and experimentation pipelines, and the early Azure Virtual Network stack when the platform was still known as RedDog.

High-throughput CDN performance.
Network telemetry and real-time experimentation.
VM network virtualization at datacenter scale.
Kernel-level troubleshooting and low-level networking.

What transfers to client work

Most clients do not operate at Meta scale—and should not copy architecture built for it. The transferable skill is knowing which reliability, observability, performance, capacity, and ownership practices matter now, and which platform work would be premature.

Continue the decision

Francisco's profile

Review his broader AI and engineering leadership experience.

Open →

AI infrastructure & LLM systems

Apply production-systems judgment to a current AI workflow.

Open →