Towards Painless Glucose Monitoring with Mid-Infrared Optoacoustic Spectroscopy and Machine Learning

19.08.2025, Abschlussarbeiten, Bachelor- und Masterarbeiten

Master Thesis on Machine Learning Methods for non-invasive Glucose Measurements

As part of the EU-funded GLUMON project, we are developing a novel approach for non-invasive glucose monitoring based on optoacoustic mid-infrared spectroscopy. This method has the potential to enable continuous and pain-free blood glucose measurement — a significant step forward in diabetes care. In an ongoing clinical pilot study and several subsequent validation studies, we are collecting high-dimensional spectroscopic measurement data from healthy volunteers as well as diabetic patients. These data will be analyzed using machine learning methods to investigate their correlation with simultaneously acquired reference blood glucose measurements.

Thesis Objectives

The goal of this thesis/project is to develop machine learning approaches for modeling spectroscopic data, apply them to clinical datasets, and thereby contribute to a better understanding of non-invasive glucose measurement using mid-infrared optoacoustic spectroscopy.

Tasks

• Preprocessing of measurement data
• Exploratory data analysis to identify relevant features and patterns
• Development of a machine learning pipeline in PyTorch for data modeling
• Selection, implementation, testing, and comparison of suitable model architectures

Requirements

• Enrolled Master’s student (m/f/d) in Physics, Computer Science, Data Science, Electrical Engineering, or a related field
• Strong proficiency in Python, including experience with data processing libraries (NumPy, pandas, SciPy)
• Familiarity with machine learning workflows and frameworks (preferably PyTorch)
• Understanding of statistical data analysis
• Ability to work independently and document code and results clearly

What We Offer

• Opportunity to contribute to a cutting-edge, interdisciplinary research topic with real-world impact
• Involvement in the EU GLUMON project
• Exposure to both academic research and practical applications through collaboration with a MedTech startup
• Flexible working conditions (on-site work and partial remote possible)

Interested?

Send an E-Mail including a Letter of Motivation and your latest Transcript of Records to: alexander.prebeck@tum.de

Alexander Prebeck (Chair of Biological Imaging, Institute for Biological and Medical Imaging)

TranslaTUM, Einsteinstraße 25, 81675 München

Kontakt: alexander.prebeck@tum.de