Syncro

MSPs run their entire business inside one console — devices, tickets, billing, alerting — so the platform has to hold two related products, dozens of third-party integrations, and a release process where a regression is somebody's outage. Shipping continuously into that without breaking tenants is the actual engineering problem.

We worked across a Ruby on Rails backend and a React + Vite front end living in a shared monorepo covering both products, with a Storybook-backed component library so the two stay visually and behaviourally consistent. Feature flags gate in-progress work so trunk stays releasable, and twelve staging environments let changes be validated in isolation before they reach production. Postgres on AWS through an ORM holds the data, S3 handles storage, and OAuth plus Stripe cover access and billing. Rollbar and incident alerting close the loop when something does go wrong.

Two products ship from one codebase with a shared design system, 60+ platform integrations are maintained behind stable internal boundaries, and unfinished work reaches trunk safely behind flags instead of waiting in long-lived branches.