The problem
TODO — 2–3 short paragraphs. What was broken or missing before this existed. Written for a non-technical reader.
What I built
TODO — 3–5 subsections, each with a real 2x screenshot. Decisions, not a feature list: what you chose, and what you chose it over.
The hard part
TODO — exactly one genuine technical or product challenge, and how it was solved. One. A list of three reads as a list of none.
The result
TODO — numbers where they exist, honest qualitative where they don't. Never generate a plausible-sounding metric.
What I'd do differently
TODO — keep this section. Admitting a trade-off is the cheapest trust move available, and it is what separates this from a template case study.