The foundation most of it stood on
For a fast-moving logistics operation, I built the data architecture around how the operation actually moved and the decisions teams needed daily, then layered AI applications on top only where they improved a real decision.
- Role
- Led the architecture
- Team
- The four-person team I led
- Scale
- Real time, three countries
The mess
Operational data was scattered across truck sensors, driver apps and business systems, and didn’t yet describe the operation reliably.
The question
Which decisions does the operation need to make every day, and what data model makes those decisions answerable?
The call
Model the operation around the decisions teams make every day, not around the systems the data happens to come from. Each application then became one decision on a shared foundation, instead of a new platform.
What changed
- The shared foundation under trip reconstruction, fuel detection and most of the company’s operational AI
- New real-time use cases shipped without new pipelines
Two of the systems that stood on it
“The highest-leverage AI work is often data architecture. When the foundation reflects the operation, the right AI use cases become obvious, and cheap.”
Delivered within a prior employer’s environment. Details are intentionally anonymized. I’m glad to discuss the operating problem, design principles, and production tradeoffs.
Have a workflow that looks like this?
Let’s work out whether AI is actually the right lever.