The proposed PhD program aims to develop a new computational tool for photon radiation shielding analysis, conceptually similar to widely used engineering codes such as MicroShield, but extended with modern artificial intelligence (AI) capabilities to improve usability, efficiency, and decision support. The core physical calculation method will be based on the straight-line attenuation method combined with buildup factors, which provides a practical and computationally efficient framework for shielding design and dose rate estimation in radiation protection and nuclear engineering applications.

The scientific objective of the project is twofold. First, the PhD student will develop a reliable and transparent calculation engine for radiation transport in simpler and moderately complex shielding configurations, including point, line, area, and volume source models where appropriate. The code will incorporate photon attenuation, material-dependent absorption properties, and buildup factor corrections for scattered radiation. Special emphasis will be placed on numerical robustness, verification against benchmark problems, and comparison with established shielding software and reference calculations. Buildup correction factors will be determined on a large variety of shielding arrangements, using a large number of Monte Carlo calculations, the results of which will be processed using machine learning algorithms.

Second, the project will integrate AI-based assistance functions into the code in order to create a more intelligent and user-friendly engineering tool. These AI algorithms will not replace the underlying physics-based calculations; instead, they will support the user in model creation, input checking, material selection, geometry setup, interpretation of results, and identification of possible errors or inconsistencies. For example, AI could help recommend shielding materials, detect unrealistic source or geometry parameters, propose optimized calculation settings, explain output trends, and guide less experienced users during problem setup.

The PhD work will include the following major research tasks:
(1) formulation and implementation of the physical and numerical basis of the shielding model;
(2) development of the software architecture and user-oriented computational workflow;
(3) design and training of AI-supported modules for guidance, optimization, and error detection;
(4) validation of the code through analytical test cases, benchmark comparisons, and selected practical applications;
(5) assessment of the reliability, limitations, and engineering applicability of the developed tool.

The expected outcome is a new generation hybrid computational tool, combining deterministic radiation physics with intelligent user support. The code will be scientifically sound, computationally efficient, and significantly easier to use than conventional shielding calculation tools. The research is expected to contribute both to radiation shielding methodology and to the broader field of AI-assisted scientific and engineering software development. In practical terms, the developed tool could support shielding design, radiation safety evaluations, educational applications, and preliminary engineering studies in nuclear facilities, medical radiation environments, radioactive waste management, and related areas. Depending on the parallel software developments, the code, as a product, can be sold or offered as open source.