I build interfaces
that ship.

Pulse pulls engineering metrics from Jira and GitHub into one live dashboard. As the data and number of views grew, the frontend was firing redundant API calls and re-rendering heavily, so it felt sluggish exactly when teams leaned on it. Non-technical stakeholders also couldn't get an answer without asking an engineer to pull it.

I built a React Query caching layer with selective context splitting and targeted memoisation, so each view only fetches and re-renders what actually changed. On top of that I built Pulse AI, a natural-language query interface that lets anyone ask for a metric in plain English instead of navigating the dashboard.

Caching and context splitting add real complexity: cache invalidation and a more fragmented state tree are harder to reason about than fetch-on-render. I accepted that overhead because the responsiveness gains were measurable and the dashboard is used every day.

Cut redundant API calls by ~50% and reduced re-renders by ~45%, making the dashboard noticeably more responsive at scale. Pulse AI gave non-technical stakeholders self-serve access to project and developer insights without writing a single query.