Senior Data Scientist · Applied ML Engineer

Ali Abul-Hawa, PhD

Selected work across multilingual retrieval, large-scale Arabic NLP, visual anomaly detection, local document Q&A, reliable software systems, and earlier control/energy research.

Arabic NLP at scale700M+ raw social-media records with streaming ingestion, MongoDB deduplication, source filtering and Arabic preprocessing
Industrial anomaly detectionDINOv2 reached 0.941 mean image F1 across 15 MVTec AD categories versus 0.929 for PatchCore; DINOv3 achieved the best AUPIMO at 0.680
Semantic model selection9 embedding variants evaluated across 28,400 documented comparisons, with MPNet selected for the multilingual retrieval system
100-class computer visionVGG16 transfer learning on UECFOOD100 reached about 70% top-1 and 92% top-5 accuracy

Selected work

Systems, analyses, and technical case studies

Independent engineering

Software projects

Earlier work & research

Computer vision, control, and energy systems

Background

Engineering foundations, applied ML systems

Electrical-engineering research in dynamic systems and control was followed by large-scale Arabic NLP, multilingual retrieval, applied machine learning, local LLM systems, and reliability-focused software engineering.