Joshua Holstein

Background:
The current era of digital transformation has led to an increase in the complexity and intricacy of manuals for machinery and equipment, particularly in the manufacturing sector. As machines continue to evolve, the necessary documentation to operate and troubleshoot them becomes more extensive. Nevertheless, employees may find it tedious and time-consuming to navigate through these voluminous manuals during breakdown situations. Traditional search methods often yield results that are limited in scope, potentially overlooking crucial contextual nuances. With their understanding of context, semantics, and natural language semantics, Large Language Models (LLMs) have the potential to revolutionize the search process by providing more efficient and targeted results.

Research Goal:
In partnership with Bayer AG, the aim of this thesis project is to leverage the capacities of large language models to create a neural search system for manuals of manufacturing machines. Our key aim is to assist workers in promptly identifying relevant solutions in the event of machine breakdowns or operational challenges. This will include fine-tuning the LLM on various manuals to understand specific terminology, common problems and their solutions. Employee feedback and experiences will be collected through interviews to fine-tune the neural search to meet real-world requirements. Following development, the neural search model will be integrated into a user-friendly interface and tested in real-world situations for effectiveness and accuracy.

In the course of this thesis, you will:

- Apply your theoretical understanding to address real-world manufacturing problems

- Explore advanced techniques such as large-language models

- Collaborate closely with domain experts in the effort to enhance the efficiency of manufacturing processes

We invite applications from those who:

- Have a strong interest in machine learning and data analysis

- Are self-driven and goal-oriented, eager to address contemporary real-world problems, and motivated to contribute original ideas

- Possess good English language skills, as the thesis will be written in English

Details:

Start: Immediately

Duration: 6 months

Salary: Yes

Location: Bayer AG, Remote or nearby Grenzach-Whylen

We promise a challenging research topic, attentive mentorship, and an environment that fosters both practical and theoretical skills. Interested candidates should send their CV, transcript of records, and a brief letter of motivation to Joshua.Holstein@kit.edu.