Manuel Ramos Calderón

Making more precise, faster and easier R&D processes in volved in the design of new BASF products through the introduction of computer vision and advanced image analysis into related workflows. Worked on the improvement of the opera tive of UAVs for data acquisition and processing.

• Model training, software design, implementation and deploy ment of a national train infrastructure monitoring system that works in real-time. Optimized models and processing pipeline to meet tight stability and run-time performance restrictions.

• Concept design, infraestructure setup, research and implemen tation of a computer vision platform for automatic analysis and inspection of photovoltaic plants using UAVs thermal and visual RGB images.

• Geolocalization of thermal defects via the geoprojection of imaging data onto the Earth surface. Geographical Information Systems (GIS).

• Model development and pipeline deployment using cloud services (Azure).

• Participated in R&D projects using image classification, object detection, keypoint detection, multiple object tracking and person re-identification.

• Leveraged the Django web development framework for model deployment into production. Designed and implemented front/back ends. RESTful APIs and real-time streaming via websockets.

• Performance analysis and benchmark evaluation of deep learn ing models. Model optimization via quantization.

• Camera calibration for photogrammetry applications.

• Technical project drafting (work packages design, resources estimation and time planning). Directed masters thesis projects for computer vision students.