GEPARD – Assistance System with Generative AI for Efficient Production
The research project GEPARD develops an innovative assistance system based on generative artificial intelligence (GenAI) that supports companies in creating and adapting production-relevant documents and data significantly faster and more efficiently. Particularly during product variant changes, production ramp-ups, or minor product modifications, high effort is currently required, as work plans, assembly plans, CAM data, or other documentation must be manually recreated or revised - often involving extensive expertise and repetitive work.
Objectives and Added Value
The GEPARD assistance system partially or fully automates this process, enabling:
- More efficient use of skilled personnel: Routine tasks are reduced, allowing employees to apply their expertise more effectively.
- Knowledge preservation: Implicit experiential knowledge from production is systematically captured and made usable.
- Improved responsiveness: Companies can react more quickly to customer requests or change requirements.
- Increased productivity and flexibility: Industrial production becomes more resilient and adaptable.
Project Content
Within GEPARD, concrete results are being developed, including:
- Prototype of a GenAI assistance system: Development, testing, and integration of a functional prototype (MVP) for at least two practical use cases.
- Modules for data and document processing: Concepts and implementations that integrate company data (such as CAM or SAP data) and domain-specific inputs into the system.
- Methods for transferability: Development of processes, guidelines, and best-practice approaches that facilitate company-specific customization and system integration.
Innovation and Application
A special focus of the project lies in the domain-specific use of generative AI in production - going far beyond generic language or AI tools. GEPARD integrates heterogeneous and unstructured data sources and transforms long-standing production knowledge into an adaptive assistance system that can be specifically embedded into industrial workflows.
Consortium & Funding
_.jpg)
The project is funded by the Federal Ministry of Research, Technology and Space.
Funding Agency: DLR Funding Agency
Project duration: February 1, 2025 – January 31, 2028
Project Consortium & Associated Partners
Related videos
Your Project Team
Tim Schroeder
Tim Schroeder studied Electrical Engineering & Information Technology as well as Business Administration and General Management at RWTH Aachen University. As the Head of Artificial Intelligence at the INC Innovation Center, he led multiple bilateral and multilateral projects in the areas of Industry 4.0, Logistics 4.0, and Artificial Intelligence. Through technology scouting, data analysis, and AI assessments, he assists companies in successfully implementing new innovations.
Tim Schroeder
Dr. Christoph Rippe
Dr. Christoph Rippe studied business mathematics with a focus on stochastics and optimization at Otto-von-Guericke University Magdeburg and then did his PhD in operations management on inventory management problems for service technicians. At INC Innovation Center, Christoph is a Senior Technology Specialist responsible for data analysis and prototypical implementation of AI use cases. He also supervises trainings and works on research topics related to AI applicability (small data augmentation/transfer learning).
Dr. Christoph Rippe
Dominik Joosten
Dominik Joosten studied electrical engineering at Hochschule Niederrhein University of Applied Sciences and completed his master’s degree in automation technology at RWTH Aachen University, including an Erasmus term focused on machine learning at NTNU in Norway. At INC Innovation Center, he is a Senior Technology Specialist responsible for developing AI solutions for industrial applications, building scalable cloud‑based systems, and driving MLOps practices. He also conducts workshops to strengthen internal AI competencies and works on data‑driven process optimization, predictive quality in manufacturing, and the practical use of GenAI technologies such as LLMs and RAG.
.jpg){kind=spacer}
%20(5).jpg)
%20(8).jpg)
