Connected ERP platform
Vividia Suite
Ten business applications, from accounting and HR to healthcare and manufacturing, connected by one sign-in, shared permissions and a common data platform. I own the frontends of five of them.
- Role
- Senior Software Engineer, frontend owner and full-stack contributor
- When
- 2024 to now
- 10
- apps
- 14
- NestJS services
- 5
- frontends I own
The problem
Accounting, HR, school records, inventory and the rest usually live in separate tools. Vividia Suite puts them on one platform: each application can stand alone while sharing identity, roles and operational data.
What I built
- Workspace, the sign-in and organization hub every other app hangs off: shared JWT, workspace switching, roles and permissions, with 300+ commits from me.
- The HRM frontend: employee records, onboarding, attendance and shifts, leave, payroll and self-service. I'm its top contributor with 370+ commits.
- The Learning Management frontend with 120+ page routes, plus backend work on its video pipeline.
- The Clinic app, backend and frontend, built from nothing: appointments, live queues, consultations and billing.
- The Superadmin console that provisions and manages organizations across the whole suite.
One identity layer across fourteen services
The backend is made up of 14 NestJS services on PostgreSQL, with a schema per service and organization-level access on every query. Frontends read the same JWT and workspace context, so a user moves between HRM, accounting and inventory without signing in again, and a role granted in Workspace applies everywhere.
Frontends that survive enterprise data
ERP screens are mostly dense tables, long forms and approval flows. Across the apps I own, server state lives in TanStack Query, forms are built with React Hook Form and Zod, and routing respects permissions, so each new module reuses the same working patterns.
Video at scale for Learning Management
Course video uploads go straight to S3 with presigned URLs. A BullMQ worker on Redis transcodes them to HLS with FFmpeg, and Socket.io pushes progress to the browser, so large lecture files never pass through the API server.