Skip to content
WorkIndofood KSA — Internal Management System

IMS

IMS is an internal management system for manufacturing operations at Indofood KSA. I handled the AI side end to end — taking raw operational data through cleaning and preparation, then serving the resulting model through an API the main application calls.

Role
AI Engineer
Discipline
Manufacturing, AI
IMS — the shift overview, demand forecast and downtime root-cause log

Built with

  • Python
  • Machine learning
  • REST API
  • Enterprise data integration

Manufacturing operations run against company standards, and a model that quietly disagrees with those standards is worse than no model at all.

The design stays deterministic first: rules produce the baseline, model output comes second. That ordering means results stay consistent with company standards and can be audited afterwards — you can point at any output and say which layer produced it and why.

01

Data preparation

Cleaning and structuring operational data from enterprise systems into something a model could actually learn from.

02

Model serving

Exposing the trained model through a REST API, so the main application consumes predictions like any other service.

03

Fitting existing screens

Worked with the fullstack team to place model output inside interfaces staff already knew, rather than asking them to learn a new one.

Next project

SIRANA

SiRANA — the real-time emergency map, on desktop