# Lead-time prediction for machined parts

> A machine-learning model that predicts the finish date and time of a machined part from the moment the ERP declares it to be made.

HTML version: https://sean.lamet.dekrow.com/work/safran-predictive-maintenance/

- Period: Apprenticeship, 2024–present
- Role: Apprentice, Industrial IT Innovation
- Status: In development
- Stack: 

## Problem

Once the ERP declares a part to be made, the people who plan production need to know when it will be finished: the date and the time. How long each operation takes depends on the load on the machines and on which operators are available.

## What I built

During my apprenticeship at Safran Landing Systems, I am building a machine-learning model that predicts the finish date and time of a part from the moment the ERP declares it to be made. It learns operation times against machine load and operator availability.

I am building it with my department, which knows the network, the environment and the services it can reach; with operators, for the features and screens they need and the data they produce; with production planners; and with the operators' managers.

## Where it stands

It is in development: I am still building it, and it is not in production. This work is confidential, so this page names no internal system and shows no architecture, URL or screenshot.