Ayoub BOUSNANE

Optimization of the machine learning demand forecasting pipeline for Decathlon Digital's Demand & Assortment Planning team. Modernized ETL data processing stages and accelerated deep learning model training/inference in an HPC/cloud environment (Databricks, AWS, GPU clusters). Key realisations: - Refactored data processing pipeline using Polars (CPU) and RAPIDS cuDF (GPU) achieving 3.6x-20x speedup - Implemented mixed-precision training (AMP-TF32) and optimized inference with torch.compile (FP16) for 11.5%-53.7% performance gains - Reduced cloud infrastructure costs by up to 90% while preserving forecast accuracy - Supported Decathlon's IT sustainability initiatives through compute optimization

During my final year at ESI, I designed and developed an API Marketplace platform enabling interactions between API providers and clients. Key contributions: - Implemented dashboards for API providers and clients to monitor activity - Processed real API usage data to generate statistics and insights for clients - Built a secure billing system supporting multiple payment options - Performed security testing and API vulnerability assessments * Technologies used: NextJs, NestJs, GoLang, Node.js, Stripe

- Designed and developed key components of the "Système de Gestion de Base de Vie" (Base Life Management System) to improve the operational efficiency. - Have Collaborated with a cross-functional team to gather requirements, design the system architecture, and create user-friendly front-end interfaces. - Have Used modern web technologies to enhance the user experience, to ensure responsiveness and accessible design such as reactjs, and its rich UI libraries (material UI) . - Contributed to optimizing workflows within the information system, focusing on data management and user interaction. - Have Gained valuable experience in working within a corporate environment and applying software engineering best practices.