Industrial AI
Models under operational constraints.Failure detection, decision-aware model selection, industrial monitoring, and deployment-oriented evaluation.
01 / RESEARCHERULSAN · KOREA
TRUSTWORTHY INTELLIGENCE FOR THE PHYSICAL WORLD
AI that remains useful when reality disagrees.
I am Alireza Mahvelati Shamsabadi, a PhD candidate at the University of Ulsan working across Industrial AI, trustworthy machine learning, physical systems, and environmental engineering.
Accurate prediction is not enough. I study when models fail at the decision layer — and how AI can remain useful under physical constraints, uncertainty, drift, and limited operational resources.
AREAS OF WORK
One research identity. Multiple physical domains.
AI for Physical Systems
Learning that respects dynamics.State, disturbances, controls, constraints, uncertainty, and interaction between models and physical systems.
Environmental AI
Intelligence for resilient infrastructure.Urban water systems, wastewater treatment, climate resilience, environmental monitoring, and operational early warning.
Scientific ML
Prediction structured by science.Machine learning informed by physical structure, scientific constraints, uncertainty, and mechanistic reasoning.