Mustafa
Awd
Contact
Areas of expertise
Biography
I am Dr.-Ing. Mustafa Awd, Assistant Professor of Mechanical Engineering and Director of the Process Informatics Group at the University of Turku, Finland. I earned my Dr.-Ing. degree in Industrial Machine Learning from TU Dortmund University (2022), M.Sc. in Manufacturing Technology from TU Dortmund (2017), and B.Sc. in Design & Production Engineering from the German University in Cairo (2014). Before joining the University of Turku, I served as a Senior Mechanistic Machine Learning Scientist at the Leibniz Institute for Materials Engineering (Leibniz-IWT) and a Mechanistic Machine Learning Scientist at Testia GmbH (Airbus Group), as well as at the Institute of Informatics and Automation of Bremen City University of Applied Sciences.
Teaching
My teaching philosophy centers on preparing students to drive innovation in AI-enhanced manufacturing through adaptive, level-appropriate mentorship. I employ evidence-based pedagogical methods, including active learning, flipped classrooms, and project-based instruction across courses in AI for Industrial Applications, Computational Solid Mechanics, and Data Science.
I tailor supervision strategies to each academic level: providing structured frameworks and scaffolded learning for bachelor students, fostering independent problem-solving and critical analysis for master students, and adopting a candidate-focused mentorship model for doctoral candidates that balances research excellence with career-specific professional development. My commitment to student success is evidenced by positive evaluations highlighting my ability to explain complex concepts clearly, my integration of real-world industrial case studies from partners, and my use of modern tools (GitHub, Jupyter, Moodle) that mirror professional AI engineering workflows.
Research
My research transforms industrial manufacturing through explainable AI and mechanistic machine learning approaches that couple physics-based models with data-driven algorithms. As Director of the Process Informatics Group at the University of Turku, I have developed interpretable digital twins for additive manufacturing that integrate quantum mechanical principles from atomistic simulations (VASP, CP2K) with physics-informed neural networks, achieving unprecedented accuracy in very high-cycle fatigue prediction.
My work pursues three interlinked streams: (1) developing explainable AI frameworks that incorporate physically meaningful constraints and multi-scale modeling to ensure interpretability in safety-critical applications; (2) creating real-time process monitoring and adaptive control systems using ensemble models with Bayesian uncertainty quantification, achieving up to 30% defect reduction in industrial partnerships; and (3) designing federated learning architectures that enable privacy-preserving collaborative learning across manufacturing sites through differential privacy mechanisms and personalized meta-learning strategies. I have successfully translated fundamental research into deployable solutions that measurably improve product quality and process efficiency while advancing responsible AI development that prioritizes transparency, fairness, and sustainable industrial transformation.