Dr. Shadi Sharba

• Built agentic applications with Pydantic AI and Cursor that coordinate multiple tools • Applied prompt and context engineering techniques to make AI coding assistants reliable for production oriented tasks • Automated simulation pipelines, reducing runtime by 90% • Achieved 100 times acceleration in model execution using custom machine learning methods • Managed projects and prepared technical specifications in collaboration with stakeholders

• Built predictive numerical model for nonlinear composite material behavior • Streamlined object detection in 2D/3D images for synthetic replica generation • Developed a machine learning homogenization scheme for nonlinear materials • Enabled real-time simulations by cutting runtime 99% with sparse dictionary learning • Enhanced functionality of commercial software via mixed-language programming • Mentored and trained students in data processing courses

• Developed a predictive maintenance framework 100 times faster than standard solutions • Accelerated fatigue simulations via machine learning techniques with less than 2% prediction error • Conducted research as visiting scholar at ENS École normale supérieure Paris-Saclay)

• Improved dataset processing efficiency by 20% by modernizing software packages for heterogenous data parsing

Thesis: A Machine learning Model for Structural Life Assessment

Thesis: Data Supported Time Integration for Nonlinear Systems