Built

A browser-based tool that cross-references device inventory across multiple management platforms to catch devices that fell out of sync, went stale, or dropped off monitoring. The tool matches records with a serial-first engine that falls back to hostname and friendly-name bridges, then flags mismatches, stale check-ins, missing endpoints, and naming inconsistencies for a technician to review each month.

A scheduled automation that pulls per-device values from one cloud platform's REST API and writes them into the matching record in another, so technicians can work from one system instead of two.

Most of the engineering went into the parts the prebuilt connectors don't cover. The flow calls both APIs through raw HTTP actions, handles a token-based authentication exchange with credentials held outside the flow definition, and parses paginated JSON responses into a shape it can match against. Matching runs on hostname with normalization for the naming drift between the two systems, and anything the flow can't resolve gets logged rather than matched on a guess. A tenant map drives per-tenant routing, so one flow definition serves every tenant instead of a copy per client.

A live operations dashboard built for a 4K office display, pulling real-time data from a service management platform and a VoIP phone system to surface SLA compliance, technician performance, project status, and ticket queue depth.

Originally a proof of concept, the project grew into a production system with a Dockerized deployment on network-attached storage and automated pull-and-restart deployment on every code push. Ongoing work includes an interactive version history modal, styled to match the rest of the design system, so viewers can see what changed and when.