firas zarai

• Designed and implemented a data bus for Orange Wholesale that enhanced data accessibility for more than 30 business collaborators through over 20 use cases. • Enhanced decision-making processes by exposing more than 1 billion network and cloud infrastructure logs. • Centralized data storage using Kafka and Object Storage S3. • Created and managed scalable ETL pipelines with Spark Streaming, handling data quality through cleansing,deduplication, filtering, and enrichment.

• Developed Azure Databricks notebooks with PySpark to ingest, clean, and aggregate large-scale data, optimizing it for Up Coop’s integrated solutions (cards, vouchers, web platforms) that enhanced users’ daily experiences. • Orchestrated data workflows with Azure Data Factory, managed dependencies between activities, and scheduled automatic executions. • Applied a Lakehouse architecture with Delta Lake to optimize incremental loads and manage data versioning.

• Integrated artificial intelligence into the BIM (Building information modeling) cycle to monitor safety conditions in construction sites through pictures. • Implemented an end-to-end ML cycle encompassing data labeling, computer vision model development, training, testing, and deployment. • Developed a web application using React.js and Django to serve the AI model.

Developed a tool that extracts invoice metadata from invoices images regardless of their quality and orientation.The tool is based on image processing, feature extraction and computer vision.