AMP
A real-time Manufacturing Execution System that gives SME factories what enterprise plants take for granted: live machine visibility, honest OEE and operational insight they can act on.
- Founder & sole engineer
- SaaS platform
- Manufacturing / MES
- Private demo on request
SME factories run blind.
Downtime gets logged on paper, if at all. OEE is a spreadsheet reconstructed days after the fact. Machine data exists — in PLCs, sensors and controllers — but never reaches the people making decisions. Enterprise MES suites solve this at a price and complexity small manufacturers cannot justify.
One live system of record.
AMP connects directly to machines over MQTT and turns raw telemetry into live dashboards: machine states, downtime with causes, OEE computed from real events, production planning and maintenance — with an analysis layer that flags anomalies before they become lost shifts.
Shop floor to browser, in under a second.
Edge
Machines publish telemetry over MQTT — cycle counts, states and alarms from PLCs and sensors on the shop floor.
Backend
FastAPI ingests the stream, computes OEE, downtime and production KPIs, and persists everything in PostgreSQL.
Realtime
WebSockets push live state to every connected dashboard — no polling, no refresh, sub-second updates.
Intelligence
An analysis layer surfaces anomalies, downtime patterns and shift-level insights operators can act on.
Built for the people on the floor.
The hard parts.
Unreliable factory networks
Shop-floor connectivity drops. The ingestion layer buffers and reconciles MQTT messages so KPIs stay correct after an outage.
OEE operators trust
OEE is only useful if the underlying states are right. AMP models machine states explicitly and keeps every number traceable to raw events.
Real-time at SME cost
No Kafka clusters or enterprise licences — a lean FastAPI, PostgreSQL and WebSocket stack keeps it affordable for small factories.
See AMP running on live data.
Private product demo available on request — happy to walk through the architecture too.