Amine GHERBI

Working with companies, startups and charities to implement customized data solutions that are best aligned with their business goals. 𝗖𝗹𝗶𝗲𝗻𝘁𝘀 : Brut, Total Energies, Pernod Ricard, Sofinco ...

Optimization and Parallelization of Deep Learning Applications on HPC clusters: - Scaling Deep Learning (UNet) applications on multi-GPU in a HPC environment. - Optimizing the training time of Deep Learning models. - Implementation of Deep Learning models to Super-Resolution problems applied to seismic images. - Deep Learning inference optimization (Tensorrt, ONNX) - Coauthored 3 research abstracts on scaling Deep Learning applications in Geoscience. (EAGE, SEG) Supervisors: Mauricio Araya, Denis Akhiyarov

Working in the Accelerateur team: - Implementing deep learning algorithms for semantic analysis of customer feedback. (Keras, AWS sagemaker) - Lawyer Recommender system for the legal protection department of Axa. - Driving actionable insights from data and present them to external and internal clients. (Pyspark, sklearn)

Automatic speech emphasis detection using neural networks

Network inference methods to reconstruct dynamic financial networks