Flowics
Software Engineer II, 2021 – 2026
Flowics is a live broadcast graphics platform. I was a contributor to its graphics editor (React + TypeScript + Node.js) and its web dashboard (React + Rails).
Tickers pick up where they left off
Shipped "tally memory" for broadcast crawlers: server-side Firebase Realtime Database tracking of what actually aired, letting tickers resume from the last shown item after a restart or republish; consolidated realtime events into a single connection-summary payload. This also enabled us to later on ship a feature allowing users to access listen for changes through a websocket connection.
- Firebase Realtime Database
- Node.js
- React
- WebSockets
External data de-normalized, merged into live graphics
To optimize payload size of data coming from the server, I rebuilt the data-joins engine that merges external connector data into graphics schemas — pluggable insertion/not-found strategies, coalesce operator, multi-argument JSON Pointer templates — validated against 8k-line production sports-data fixtures and delivered end to end with a server-side data migration and i18n in three languages.
- TypeScript
- JSON Pointer
- Data migrations
- i18n
One request for many fonts
Designed a bulk font-upload REST API using MongoDB transactions and optimistic concurrency (if-unmodified-since semantics), collapsing N HTTP round-trips into one and surfacing concurrent-edit conflicts as explicit per-item errors; paired it with a self-serve upload UI (including client-side font parsing and staged rollout) that replaced a support-only workflow.
- MongoDB transactions
- REST
- Optimistic concurrency
- Font parsing


