yesmine lajmi cherif

• Developed and integrated a deep learning segmentation model for industrial imagery, enhancing model performance. • Achieved a +10% improvement in Intersection over Union (IoU) and +12% in CLDice compared to existing solutions. • Contributed to an innovative project leading to a patent application, showcasing creativity and technical expertise. • Built end-to-end pipelines for data preprocessing, training, and evaluation, streamlining the machine learning workflow.

Developed high-resolution satellite image generation models using generative AI, exploring Vision Transformers and diffusion models separately.

• Coached students in C++ and algorithmic problem-solving for national-level programming competitions. • Strengthened students’ skills in data structures, complexity, and competitive programming strategies. • Developed tailored lesson plans to enhance students' understanding and performance in competitive environments.