Work record — Ross Video

Digital twin reporting

A read-only copy of the plant's data, and the dashboards on top of it — so anyone could ask the production systems anything, without asking anything of the production systems.

Role
Senior developer — solo on this project
Timeline
2020 — 2025
Stack
Angular · TypeScript · C#/.NET · SQL Server · Grafana
Scope
Primary internal tooling for quality, manufacturing, logistics, admin, and reporting

Behind the firewall

These systems are internal to Ross Video — their code and screens aren't mine to publish. This record documents the problem, the design, and the outcome.

The problem

The systems that run a plant — ERP, MES, test, logistics — are terrible places to run reports. Every ad-hoc query competes with the transactions keeping production moving, so reporting gets rationed: locked down, batched overnight, or funneled through whoever has database access. The people with questions about production data were exactly the people who couldn't touch it.

What I built

Decisions that mattered

Separating "the data you run the plant on" from "the data you ask questions of" is the whole design. Once the reporting copy existed, saying yes to a new dashboard or an unusual query became cheap and safe — the failure mode of a bad query changed from "slowed down production" to "slowed down a report."

The twin stayed current through scheduled query jobs — plain SQL on a schedule pulling from the production systems, not an exotic replication stack. Boring on purpose: easy to reason about, easy to extend when a new dashboard needed data the twin didn't have yet.

The reusable module library was a bet that there would always be a next internal app. There always was. Paying a design tax on the first app to make the fifth one fast is the kind of decision that never shows up in a sprint demo and dominates the economics of internal tooling.

Outcome

Operations staff across quality, manufacturing, and logistics worked in these applications daily for years. The digital twin itself was mine end to end — the scheduled jobs that kept it fed, the queries behind the reports, and the Grafana dashboards on top — designed, built, and run solo.