Learning firmware from the build outward

An unfamiliar firmware codebase rarely explains itself in one place. The useful picture is distributed across the source, its build configuration, the artifacts it produces, and finally the behavior of the hardware.

During roughly twelve weeks at Coherent, I compared embedded development environments and checked whether the same firmware continued to behave as expected. That made migration differences concrete enough to investigate and test.

The artifacts are part of the explanation

The build output is not just something produced at the end. It is evidence. Comparing what each environment produces with what happens on hardware turns a vague migration problem into specific questions that can be tested.

My part of the work was to make that investigation repeatable: create clear validation steps, document the findings, and leave behind a process another engineer could continue using.