Big tech · Big tech
Making misattributed assets visible across a ~700K-server fleet
Financial attribution for private cloud infrastructure that runs internal AI workloads, plus internal platforms that make livability, deployment state, and cost visible to the people who own them.
Headline
~700K
servers under financial attribution
Private cloud
for internal AI workloads
Weekly
internal platform used by hundreds of engineers
Millions
in budget distortion from quiet misattribution
The problem
At fleet scale, a small rate of misattributed physical assets doesn’t look like an error. It looks like a budget. Across roughly 700,000 servers, those quiet mistakes distort planning by millions of dollars, and because nobody can see them, nobody owns them.
The work
I worked on fleet-scale private cloud infrastructure for internal AI workloads. Part of that was the attribution pipeline that corrected misattribution, turning an invisible accounting problem into something leadership could see and act on. Part of it was internal platforms, used weekly by hundreds of engineers, that gave visibility into private cloud livability, deployment state, and the financial attribution of physical assets.
The design principle
Build finished product experiences, not dashboards someone duct-taped together. Financial admin, observability, livability tooling: each one designed so the people who have to live with the data actually use it.
Details are limited to what’s public about the role, and nothing confidential is shared here.