Vacancies
Forecasting energy smarter with AI at Gasunie
Data Science & AI

The challenge

Within Gasunie, several machine learning models were ready in the form of Proof of Concepts, such as:
  • FlowForecaster Model (gas flow prediction).
  • Energy Procurement Model
  • Various compliance models related to actuators
  • Models had to be made production-ready (MLOps, CI/CD)
  • Models had to be transparent to regulators (explainability)
  • Models had to be fit within existing infrastructure (Azure, DevOps)

At the same time, a new model for regulatory controls also had to be developed.

Our approach

Our consultant worked with various business units and IT teams to restructure, validate and put models into production. Key components of the approach:
  • MLflow and Azure ML Studio implemented for model tracking and experiment management
  • Explainable AI (SHAP) deployed for transparency towards stakeholders
  • CI/CD pipelines built in Azure DevOps
  • Streamlined processes around troubleshooting and environment management
  • R and Python combined for powerful reporting and predictive modeling
  • Intensive collaboration with stakeholders for domain knowledge translation

context

Gasunie manages the transport network for gas in the Netherlands and northwestern Europe. In a rapidly changing energy market – with pressures on sustainability, security of supply and international dependencies – data-driven optimization plays a key role. IBS provided a Machine Learning Engineer / Data Scientist who helped Gasunie operationalize their energy prediction models as well as meet increasingly stringent compliance requirements.

The impact

  • FlowForecaster and Energy Procurement Models now scalable and maintainable
  • Regulator model operational with full explainability
  • MLOps best practices embedded in the team
  • Data science has become a structural part of Gasunie’s decision-making

Why this case matters

This assignment shows that real impact of AI only occurs when models are integrable, explainable and maintainable. And that is exactly where Isatis’ added value lies – we bridge the gap between data science and software engineering.

Isatis business solutions

Whether it’s asset monitoring, energy forecasting or compliance tooling – we help companies in energy and infra with scalable AI solutions, with governance.