Master's Thesis: Leveraging LLMs to Enhance Industrial Process Efficiency

Actemium

As a part of the global VINCI Group, which employs over 200 000 people, Actemium specializes in providing solutions in industrial automation, process, and manufacturing industry, as well as energy and infrastructure. Energy AI, based in Gothenburg, is a Swedish business unit focused on advancing AI-driven solutions for industrial efficiency.

Objective

This thesis aims to integrate Large Language Models (LLMs) into industrial processes to develop a system that recognizes, tokenizes, and analyzes components using AI-driven tools. The research will focus on how LLMs can optimize workflows in image recognition and data traceability, significantly reducing development time and improving accuracy in industrial settings.

Scope & Outline

The project will involve applying LLMs to process data extracted from image recognition of industrial schematics, bridging these schematics with the generation of structured data. This work will culminate in a functional prototype and a comprehensive analysis of the impact of LLMs on process efficiency within industry. The thesis will be done during spring of 2025 (20 weeks) by one person in Gothenburg, Sweden.

Expected Results

• Development of a streamlined system utilizing LLMs to enhance efficiency in industrial processes.
• In-depth insights into image recognition, data tokenization, and system integration for complex industrial applications.
• A final report detailing implementation strategies, challenges, and solutions for applying LLMs in an industrial environment.

Prerequisites

• Academic background in Engineering Mathematics, Data Science, or related fields.
• Familiarity with LLMs and computer vision techniques is recommended.

Contact

Amanda Sjöland Lead Engineer – Optimization
amanda.sjoland@actemium.com
0765118006