Fehmi Ghouil

– Re-architected legacy PL/SQL batch processing into Java 17 microservices for real-time back-office derivatives system – Reduced critical processing time from 30 minutes to 3 minutes (10x faster) using parallelism, memory tuning, and query optimization – Designed concurrent pipelines with multithreading and Java Streams to handle high-volume financial transactions – Tuned complex SQL/PLSQL queries using CTEs and optimizer hints, reducing DB load – Owned end-to-end migration (design → implementation → production rollout → support) – Migrated legacy desktop screens to a modern Angular frontend, improving usability, maintainability, and user experience for back-office operations. – Provided L2/L3 production support for client-critical financial workflows

– Developed a specialized microservice using ASP.NET following Clean Architecture principles to configure, monitor, and optimize Electronic Payment Terminal (EPT) usage within a local platform. This architecture ensured separation of concerns and maintainability. – Incorporated advanced algorithms and workflows, leveraging CQRS (Command Query Responsibility Segregation) for seamless EPT integration and customization. This approach enabled efficient command and query handling, improving performance and flexibility in managing EPT configurations. – Facilitated seamless communication and data exchange with internal modules and the authorization server, utilizing robust communication protocols and secure APIs. – Implemented a desynchronization management system using Big Data techniques on Azure Databricks to ensure data integrity across distributed EPT deployments. This system processed large volumes of data in real-time, minimizing synchronization issues – Integrated alert systems within the microservice to promptly address EPT data inconsistencies. Alerts were managed through Azure SQL Server, enabling rapid responses to operational challenges and ensuring uninterrupted EPT services.

– Utilized Python's capabilities to implement Monte Carlo simulations for generating representative samples from incomplete datasets. This approach significantly reduced dataset size by 99% while preserving the probability distribution integrity. – Developed a user-friendly dashboard leveraging Jupyter Notebook and ipywidgets to facilitate the sampling process. ipywidgets were instrumental in visualizing probability distributions and providing interactive controls for analysis. – Integrated Matplotlib and Seaborn for data visualization within the dashboard, allowing for detailed analysis of the generated samples and comparison with original datasets.

– Gained hands-on experience in a professional setting, collaborating with team members and understanding workplace dynamics and project workflows. – Contributed to a Java-based project, enhancing my programming skills and understanding of software development processes, including coding, debugging, and testing.