Zakariae Mrabet

Led data migration from Openfire to Business Central, ensuring seamless transition and data integrity. Managed ETL processes, maintaining and upgrading data pipelines using Snowflake, dbt, and Fivetran, improving data flow efficiency. Collaborated with cross-functional teams to troubleshoot and resolve issues, enhancing overall project outcomes and system performance.

OCP AI Research and Development Lab • Built the CoreStream project, a scalable streaming platform supporting RTSP, HTTP, and various camera protocols with interchangeable deep learning models. • Led the EpiVision project, an AI-powered safety system detecting HSE violations (e.g., missing helmets, vests, and restricted zone breaches) and triggering automated alerts via email or access control restrictions. • Worked on the ConveyorVision & RockVision project, automating anomaly detection in conveyor belts and sieves at OCP to optimize maintenance and streamline data collection. • Responsible for deploying deep learning and machine learning models using NVIDIA Triton Inference Server, ensuring high-performance AI inference in lab environments. Stack: Python, PyTorch, YOLO, NVIDIA Triton, DeepStream, FastAPI, Django, Docker, CUDA, PostgreSQL

In my role as a Data Scientist at the Ministry of Interior in Oujda, Maroc, I developed a Decision Support Information System for Human Resources Management, enhancing decision-making and prediction for improved performance. By implementing advanced features to streamline HR processes and collaborating with cross-functional teams, I ensured system alignment with organizational goals, improving overall HR efficiency.

I enhanced decision-making by implementing advanced features for purchase order processes, leading to an 80% reduction in processing time. Collaborated with cross-functional teams to ensure seamless integration of new features, enhancing overall system performance and user satisfaction at ONSSA in Oujda, Maroc.