This Master’s thesis focuses on the analysis, modeling, and interpretation of high‑resolution climate data from forest edge environments. Using long‑term measurements of temperature, moisture, and radiation, the work aims to identify and quantify microclimatic patterns related to different forest edge structures.

How does human posture change when drill height, tool design or applied force vary? Find out – using real biomechanical data from the CoSiMMI research project! In this thesis you will develop an AI model that predicts realistic working posture form joint angles, EMG, grip forces and force-plate data. Your work will help design more ergonomic tools and explore how AI van truly understand human movement. Interested in machine learning, biomechanics and real-world data? Join us! Requirements: Interest in biomechanics, ergonomics, or AI Basic skills in Python or MATLAB First experience with machine learning is an advantage Enjoyment of data analysid and modeling The project can also be completed as a team – feel free to encourage interested friends or fellow students to join and apply together. For more insights into the project please contact me.

Im Rahmen dieser Masterarbeit soll in Kooperation mit einem Industriepartner untersucht werden, wie aus juristischen Chat-Interaktionen qualitativ hochwertige Trainings- und Benchmarkdaten für maschinelle Lernverfahren synthetisch erzeugt werden können.

Master thesis/Masterarbeit

Master thesis: Adjoint Methods for Nonlinear Dynamic Problems: Expanded vs. Non-Expanded Formulation

Externe Masterarbeit oder Forschungspraxis bei Sureel.ai .

Are you looking for a lab to do your master's thesis in SS 2026? Are you interested in method development to add to the analytical toolbox? Then, this might be a project for you!

Jointly with Occurrence AB — a dynamic startup based in Stockholm and Munich — Chair of Perception for Intelligent Systems offer several MSc thesis topics concerning human activity analysis and digital twins in real production environments.

Accurate 3D bone models are vital for Total Knee Arthroplasty. While deep learning can reconstruct 3D CTs from 2D X-rays, boundaries often remain blurry. The aim of this project is to improve the reconstruction. By developing a 3D bone segmentation model and integrating it as a differentiable, anatomy-aware loss, we will penalize structural inaccuracies. This prevents bone blending and enforces sharp 3D boundaries, which are required for precise surgical planning.

Talk2Drive: A Conversational Preference-Aligned Driving Framework Using Multi-Modal Large Language Models

We are seeking a highly motivated student to contribute to the implementation and evaluation of real-time signal processing functionalities on FPGAs.