Soulaymane Loukili

Developing a new automatic metric for measuring semantic similarity between texts, intended for use in evaluating large language models (LLMs).

• Conducted a comprehensive literature review on PINNs, specifically in the context of computational mechanics models and optimization techniques.

• Developed a theoretical framework to enhance the performance of PINNs by integrating Partial Differential Equations (PDEs) into neural network architectures.

• Designed and executed practical experiments using pre-exascale computing, demonstrating the impact of various optimization strategies on simplified use cases.

• Collaborated with a multidisciplinary team to explore potential applications of PINNs in high-performance computing environments, contributing to advancements in accurate and efficient modeling techniques.

This work contributed to the field by identifying and evaluating methodologies to optimize the performance of PINNs for more accurate computational modeling in mechanics.

Worked as a Senior Engineer in the Treasury Department at the Ministry of Economy and Finance, Morocco, specializing in managing and analyzing financial data and debt recovery for local authorities. Focused on developing and monitoring budgetary and debt recovery statistics to optimize revenue collection, improve budget management, and modernize accounting procedures, thereby enhancing overall financial efficiency.