Research Interests

My research focuses on machine learning for technical systems, with an emphasis on simulation-based methods, calibration, parameter identification, and sensor data analysis.

One of my current research areas is the calibration of cable-driven parallel robots (CDPRs). I investigate how simulations and synthetic data can be used to identify model errors and unknown system parameters and to improve the accuracy of technical systems.

Other topics I work on include:

  • Digital twins and simulation-based data generation
  • Robotics and mechatronic systems
  • Sensors and sensor data fusion
  • Edge AI and embedded systems
  • Computer vision
  • Large language models and agentic systems
  • Natural language processing
  • Robust and explainable machine learning methods
  • Human-centric AI

I am particularly interested in projects that closely connect computer science with real hardware, measurement systems, or physical processes.

ZaKI.D

At ZaKI.D, I work on transferring current AI methods into industrial applications.

Together with small and medium-sized enterprises, we develop solutions for concrete challenges involving products, production, and business processes. My responsibilities range from technical analysis and data preparation to selecting and developing suitable methods and building prototypes and demonstrators.

Depending on the project, we use classical machine learning methods, computer vision, language models, edge AI, or custom software solutions.

Interdisciplinary Projects

In addition to developing models, I work on building complete experimental systems and prototypes. These may include microcontrollers, sensors, actuators, measurement hardware, networks, backend systems, and user interfaces.

I am especially interested in projects where an algorithm ultimately has to interact with a real technical system.

Electronics and Prototyping Lab

Coming soon

A new electronics and prototyping lab for research, teaching, and student projects is currently being established at the chair.

The planned workspaces and equipment will support:

  • Electronics development and measurement technology
  • Microcontrollers and embedded systems
  • Sensors and actuators
  • Edge AI
  • Robotics
  • Rapid prototyping
  • Building custom experimental setups

More information will follow once the lab has been completed.

Theses

I am currently not offering predefined topics for Bachelor’s or Master’s theses.

If you are particularly interested in one of my research areas and have a suitable idea of your own, you are still welcome to contact me by email. Please briefly describe the topic, your academic background, and what you would like to investigate as part of the thesis.

Supervised Theses

Upcycling einer Modelleisenbahn mit intelligenter Steuerung (February 2023 · Bachelor’s)

This thesis upgrades an outdated model railway system with microcontroller-based sensors, PWM motor control, and an intelligent control system accessible via UDP and an API.

Online Marketing: Measurement Methoden, rechtliche Basis und alternative Strategien bei eingeschränktem Tracking (June 2023 · Bachelor’s)

This thesis examines marketing attribution methods, the effects of the GDPR and browser tracking protection, as well as technical and privacy-oriented alternatives when user tracking is restricted.

Digitalisierung eines robotergestützten Synthese-Workflows für ein Katalysatorscreening-Labor mit KI-basierter Syntheseroutenprädiktion (April 2024 · Master’s)

This thesis digitizes and orchestrates the workflow of an automated catalyst-screening laboratory and uses AI to select and predict suitable reagents and synthesis routes.

Frag SANA: Entwicklung und Evaluation eines KI-Assistenzsystems für Prüfungsordnungen in der Informatik (July 2025 · Bachelor’s)

This thesis develops and evaluates a RAG-based assistant that answers questions about computer science examination regulations using various configurations of LLMs, document formats, and vector databases.

Entwicklung und Evaluation eines Prototyps zur LoRa-basierten Ortung gefährdeter Personen im Wasser (July 2025 · Master’s)

This thesis develops a LoRa prototype comprising a transmitter and two receiving stations and evaluates its suitability and accuracy for locating people at risk in the water.

Leveraging Large Language Models for Bidirectional Translation Between ODRL Policies and Natural Language (July 2025 · Master’s)

This thesis investigates fine-tuned and prompt-based language models for bidirectional translation between machine-readable ODRL policies and natural-language descriptions.

Edge-First Multi-Room-Sprachsteuerung im Smart Home ohne Cloud: Architektur und prototypische Umsetzung (April 2026 · Bachelor’s)

This thesis designs, implements, and evaluates a decentralized, locally operated multi-room voice-control system using ESP32-S3 nodes, a Raspberry Pi, and an NVIDIA Jetson, with regard to privacy, offline capability, latency, and energy efficiency.

Projects

Publications (1)

2024 (1)
  • Marcinkowski, Dominic and Weis, Torben
    WoT-PL, Harnessing Large Language Models for IoT Schema Translation: A Conceptual Framework and Preliminary Findings
    In: Proceedings of the 14th International Conference on the Internet of Things (IoT 2024)
    DOI: 10.1145/3703790.3703833

Teaching

WiSe 25/26

SoSe 2025

SoSe 2024

WiSe 23/24

SoSe 2023

Dominic Marcinkowski

Research Associate

Contact

  • Phone: +49 203 379-2727
  • E-Mail: dominic.marcinkowski[@uni-due.de]

Postal address

  • Building BC, Room 309
  • Bismarckstr. 90 47057 Duisburg, Germany

Consultation hour

By appointment