Achraf khaddouri

Extracted, cleaned, and analyzed healthcare data as part of strategic projects for several university hospitals Designed interactive dashboards, reducing the analysis time for key indicators Developed automated reporting tools that streamlined decision-making for medical Contributed to the implementation of Business Intelligence solutions at 6 healthcare institutions, resulting in a 30% reduction in data processing

During this research internship, I worked on offensive language detection in Arabizi, an informal way of writing Arabic using the Latin alphabet and numbers. The study aimed to enhance AI-based moderation tools by fine-tuning BERT-based models for better accuracy in detecting harmful content. Key Responsibilities & Achievements: -Fine-tuned language models (DziriBERT, DarijaBERT, and AraBERT) for detecting offensive language in - Arabizi. -Conducted a comparative analysis using SVM as a baseline and evaluated performance through cross-validation. -Demonstrated that pre-trained, dialect-specific models outperform generic language models. Highlighted the role of transfer learning in improving offensive language detection across Arabic dialects.

During my final-year internship, I designed and developed an intelligent intervention management application for the ONEE - Water Branch Datacenter. Key Responsibilities & Achievements: -Analyzed business requirements and prepared a functional specification document - Developed a desktop and mobile application using React Native and Flask - Integrated AI features via Microsoft Azure for image analysis and anomaly detection - Automated incident reporting and access management processes - Implemented a notification system and report generation to enhance intervention traceability - Deployed and tested the application for performance and security