Software Engineering / AI Engineering

Work Schedule

Standard (Mon-Fri)

Environmental Conditions

Office

Graduation assignment: Data-assisted diagnosis and predictive maintenance

Level: HBO bachelor or WO bachelor/master graduation project

Relevant studies: Software Engineering

Location: Thermo Fisher Scientific, Eindhoven R&D, Transmission Electron Microscopy (TEM)

Introduction

Thermo Fisher Scientific develops technologies and services that help customers make the world healthier, cleaner, and safer. Eindhoven is its primary technology location for Transmission Electron Microscopy (TEM), bringing together R&D, manufacturing, and service expertise. Thermo Fisher Scientific in Eindhoven

TEM instruments use electrons to reveal structures at extremely small scales, supporting research in life sciences and materials science. This assignment explores how software can help engineers diagnose problems in these complex instruments more efficiently.

Background and objective

Service engineers use system logs, bug reports, calibration files, and technical procedures to investigate failures. This information is often spread across multiple sources, making diagnosis time-consuming and dependent on individual experience.

The objective is to develop and evaluate a software prototype that combines historical system data with service knowledge to suggest likely causes and useful diagnostic checks.

Research question

How can historical system data and service knowledge support faster and more reliable diagnosis of a selected TEM system problem?

The student will investigate which data sources are useful, compare suitable technical approaches, and measure the prototype’s value against the current diagnostic process.

Scope and approach

The project focuses on one failure mode or a closely related group of problems, such as an offline module or recurring performance issue.

The student will:

• Analyse the current diagnostic process with service engineers.
• Assess and prepare relevant data and technical documentation.
• Compare approaches such as structured search, diagnostic rules, and an LLM-based assistant.
• Build a prototype that provides diagnostic suggestions with supporting evidence.
• Evaluate diagnostic accuracy, diagnosis time, and usability using historical cases and engineer feedback.

The prototype is intended for internal evaluation. Predictive maintenance may be explored as an extension if sufficient degradation or failure data is available.

Deliverables

• A research plan and assessment of available data.
• A working prototype with documented code and usage instructions.
• An evaluation against the current diagnostic approach.
• A graduation thesis, final presentation, and recommendations for further development.

We are looking for a student who:

• Is enrolled in an HBO or WO Software Engineering programme and is ready to start a graduation assignment.
• Has programming experience, preferably in C#. Python is also acceptable.
• Understands software design, testing, and version control.
• Can analyse data and approach technical problems systematically.
• Can work independently and collaborate with software engineers and service specialists.
• Communicates clearly in English, both verbally and in writing.

An interest in AI, machine learning, or information retrieval is welcome. Prior knowledge of electron microscopy is not required.

Apply today!