Ayoub DIOURI

• Developed a modular Python library to automate the digitization of well log rasters. • Reduced manual digitization time from 1-2 weeks to mere minutes with a fully automated pipeline. • Achieved over 93% accuracy in segmentation and 87.2% in metadata extraction using advanced visual language models. • Contributed as the first author to a research paper presented at the ADIPEC Exhibition & Conference.

• Designed an adaptive image sampling and reconstruction system utilizing compressed sensing and deep learning techniques. • Implemented two deep learning models: NetM for mask generation and NetE for inpainting-based reconstruction. • Optimized bandwidth usage for drone-based transmissions, incorporating a real-time refinement loop. • Developed an interactive Streamlit dashboard for monitoring and visualization of the system.

• Developed a high-performance Python tool to automate the translation of legacy scripts into modern Python code. • Implemented integration tests to ensure functional parity and reliability of the new code. • Achieved up to 10× faster execution speed with cleaner, through optimized code generation and execution