OUALID ALLOUCH

Goal: Automate environmental monitoring by developing a high-performance embedded AI system for drone-based video analytics. Action: Designed an end-to-end pipeline merging Deep Learning (YOLOv8) with Classical Computer Vision (OpenCV). Architected automated data workflows for frame extraction and structured metadata generation. Engineered model optimization strategies using TensorRT and ONNX Runtime to meet strict real-time latency requirements. Value Added: Successfully deployed the system on NVIDIA Jetson (AGX/Orin) platforms within a ROS2 environment. Achieved robust, low-latency inference on complex aerial datasets, bridging the gap between R&D prototypes and edge deployment.

Goal: Digitalize and secure laboratory technician training by developing an immersive VR simulation ("MicroLab VR") for the Milliflex Oasis® system. Action: Software Engineering: Developed a complex immersive application using Unity (C#), integrating procedural constraints and state-machine management for laboratory workflows. Data-Driven Feedback: Designed an analytics pipeline to track user interactions (logs, timing, error patterns) and provide real-time performance feedback. UX & Systems: Engineered an interactive pedagogical path with contextual aid and automated scoring systems based on learning metrics. Value Added: Delivered a functional industrial training tool with a dedicated KPI Dashboard for trainee evaluation. Successfully managed the project using Agile methodologies, ensuring high-quality technical documentation and iterative client demonstrations.

Goal: Support the global climate initiative CCI Lakes by automating the monitoring of inland water bodies through large-scale satellite data analysis. Action: Cloud Computing: Developed automated analysis workflows using Google Earth Engine (GEE) and Python (JupyterHub) to process massive Sentinel-2 datasets. Image Processing: Engineered algorithms for cloud-masking and temporal compositing to generate high-quality, cloud-free mosaics. Automation: Orchestrated end-to-end remote sensing pipelines (pre-processing, spectral index calculation, and automated exports) integrating ArcGIS environments. Value Added: Significantly improved the reliability of global water body datasets. Delivered robust technical documentation and optimized processing scripts that reduced manual intervention in the R&D cycle.