Francesco Carrasso Email and Phone Number

- Coordinated a cross-functional team as Project Manager to develop and deploy a cutting-edge software service for Space Situational Awareness, resulting in the successful release of the Minimum Viable Product (MVP).- Led the design and optimisation of Machine Learning pipelines (with PyTorch, pandas, scikit-learn), ensuring the safety of satellites in collaboration with the esteemed German Aerospace Center (DLR).- Implemented and fine-tuned LSTM, Random Forest and XGBoost models to assess safety in orbit with time-series numerical data.- Conducted comprehensive data preprocessing, including data cleaning, feature selection, data balancing and Principal Component Analysis (PCA) on large-scale datasets for regression and classification tasks.- Authored technical documentations and presented findings to national entities (DLR) paving the way for future research publication.

- Developed Machine Learning pipeline for Computer Vision and object detection in space environment - Conducted extensive literature reviews on state-of-the-art object detection algorithms, enriching project insights.- Employed analytics to assess dataset quality, enhancing model robustness by analysing the Signal-to-Noise ratio in object bounding-boxes.

"A Convolutional Neural Network Model as Image Processing in Cislunar Environment"Advisor: Prof. Francesco TopputoDeveloped image processing algorithm in the environment of LUMIO mission, comparing image processing and Artificial Intelligence approaches to estimate spacecraft position with respect to the Moon.- Rendered 100.000+ high quality images of the Moon in realistic conditions to create a dataset, by using NASA SPICE for RF transformations and real illumination scenario- Applied standard image processing approaches to retrieve distance of the satellite with respect to the Moon, as eigenanalysis and edge detection- Applied a Convolutional Neural Network for the same purpose, aiming to centroid and range estimation.