Software Developer Intern

This is a short write-up of the project and what it taught me, kept at the level of engineering and design ideas rather than internal specifics.

I had done a testing term at the company before this; this was the first time I was building product-facing UI, and the first time I owned the design of something end to end.

Performance data lived in several places. An engineer doing root-cause analysis had to gather the picture by hand each time - pull a few sources, line them up, and reason across them before even starting on the actual investigation. It was slow, and the process was slightly different every time.

The platform brings the key performance indicators into a single centralized view. Time-series data is surfaced so a change over time reads immediately, and you can move from a symptom toward its cause without switching tools.

That consolidation is the whole value: less time spent assembling context, more spent on the analysis that actually matters.

I worked the UX in Figma before writing the component code - deciding what to show, what to leave out, and how to lay it out while it was still cheap to change.

Designing first meant the build was mostly execution rather than a running argument about layout, and it left something concrete to validate with the people who would use it.

The stakeholders' workflows were genuinely different - Amazon and UPS were two of the named ones, and what each team wanted to see first, and how they read a chart, did not line up. The dashboard could not be tuned to one team's mental model without becoming worse for the others.

The user-centric changes were not cosmetic. Clearer visualization cut the time stakeholders spent getting to a resolution - root-cause analysis that used to take an assembly step first was closer to immediate.

That showed up where it counts: the design work tied to a measurable lift in client retention over the period.

This is the term where design stopped being decoration to me and became an engineering discipline with measurable results. A layout decision connected directly to a business outcome, and I could point at the line.

Designing for audiences as different as Amazon and UPS also taught me that the way to serve several users at once is not to average them - it is to find the need they actually share and build for that.