Using AI to improve hierarchical demand forecasting – Hilti

 Overview

Company Description

Hilti stands for innovation and direct customer relationships. More than 34,000 employees around the world, in over 120 countries, contribute to making our customers’ work more productive, safer and more sustainable. We do this with our hardware, software and service offering. With roughly 300,000 customer contacts each day, many ideas come directly from our customers. If there is a challenge for which no Hilti solution exists, one will be developed. This is why we invest approximately 7 percent of sales each year in research and development. From product development to manufacturing, logistics, sales and services, we cover the entire value-added chain. Founded in 1941 by brothers Eugen and Martin Hilti, our company builds on strong roots and continuity. This long-term commitment has supported us in becoming a reliable partner for our customers and a trusted brand that they choose to work with. With our defined purpose of “Making Construction Better”, we are committed to developing products and solutions that drive productivity, sustainability, and safety in the construction industry with our values of integrity, courage, teamwork and commitment at the base of everything we do.

https://careers.hilti.group/en/who-we-are/

Project Description

The sales & operations planning (S&OP) process at Hilti is an integrated business management approach that aligns all functions of the organization to achieve focus and synchronization. It involves updating forecasts to create sales, production, inventory, customer backorder, and financial plans. This collaborative and cross-functional process brings together major operational departments to optimize inventory, improve forecast accuracy, and enhance supply chain management. The S&OP process is not just a supply chain initiative but a comprehensive business strategy that drives profitable growth and timely business decisions. One element of the S&OP process at Hilti is a monthly statistical forecast as a baseline forecast, enriched by market intelligence. It is carried out using historical data to calculate forecasts.

Goals of the project

Develop approach to predict demand on all product hierarchy levels and markets using state-of-the-art methods.

Deliverables

• Forecasting approach that works well on all levels of aggregation for more-month-ahead predictions.
• Approach implemented in MS Fabric platform (Python code)
• Master thesis
• In-house knowledge sharing

Knowledge

• You already have sufficient programming skills and strong mathematical and statistical knowledge.

 More information: escf@tue.nl