Gaetano Settembre

AI/ML Researcher & Engineer - Engineering Consultant

Dec 2025 - Present

TMC Italia S.p.A. (at Leonardo S.p.A. - Space Division), IT

• Development and validation of AI/ML-based on-board processing solutions for innovative satellite payloads, including end-to-end processing pipelines for multispectral and hyperspectral Earth Observation (EO) data, covering corrections, cloud detection, semantic segmentation, and object detection.
• Performance, power consumption, and timing assessment of AI models and conventional algorithms on representative flight hardware, optimizing computational efficiency and resource utilization on Unibap iX10 platforms (AMD GPU, Intel Myriad X, or Hailo-8 accelerators).
• Training, validation, and maintenance of deep learning models in HPC multi-GPU environments for Very High Resolution (VHR) instrument simulators, supporting advanced processing tasks such as image refocusing and denoising.
• Contribution to the definition of space segment and payload system requirements, focusing on EO instruments and on-board processing units to ensure alignment between mission objectives, user needs, and technical constraints.

Ph.D. in Computer Science and Mathematics

Oct 2022 - Oct 2025

University of Bari Aldo Moro, IT

Scholarship D.M. 352/22 in agreement with the company Planetek Italia S.r.l., for Curriculum 2 Mathematics (ex. MAT/08, now MATH-05/A), with theme bound by title: "Low-rank models for the analysis of Earth Observation data focusing on coastal and marine environments". The research project focuses on studying and developing low rank approximation (LRA) computational models able to integrate heterogeneous data, hyper/multispectral images, and sensor data, to better understand the environmental phenomena these data represent. Existing LRA methods are tools for the representation and analysis of multivariate large datasets used in various application areas. The project aims to construct variants of LRA using constrained optimization algorithms based on appropriate cost functions. These will allow adding ad- hoc constraints on the latent components to be extracted from Earth observation data, marine and coastal data, to improve the quality of the resulting approximation model.

M.Sc. degree (Hons.) in Data Science

Sep 2019 - Mar 2022

University of Bari Aldo Moro, IT

Data Mining, Machine Learning, Mathematics and programming for Data Science, Management of structured and unstructured data, Processing of sensitive data, Statistical modeling, Numerical methods for Data Science, Management and analysis of Big Data, Decision-making and optimization models, data-driven economic models, Sentiment Analysis, Recommender Systems.

Curricular Internship activities at Institute of Crystallography c/o CNR:

• ETL (Extract, Transform, Load) process of crystallographic data from open source databases
• Exploratory analysis of crystallographic data and non-parametric statistical tests
• Training, evaluation and tuning of Machine Learning models for crystal system determination (multi-class classification task)
• High-performance Models Deployment and development of Web Platform called CrystalMELA (Crystallography MachinE LeArning)

B.Sc. in Computer Science

Sept 2015 - Apr 2019

University of Bari Aldo Moro, IT

Problem solving and software development, Discrete mathematics, Computer architecture, Operating systems, English for IT, Programming laboratory; Algorithms and data structures, Databases, Statistics, Software engineering, Advanced programming methods, Mathematical analysis, Numerical calculation; Computer networks, Knowledge engineering, Usability and UX, Information Retrieval.

Curricular Internship activities at SER&Practices (SER&P):

• Development of 'Serious games' following the RedTeam vs. Blue Team approach to improve the skills of the SOC (Security Operation Center) and CSRIT (Computer Security Incident Response Team) team
• Study of malicious attack scenarios on Linux Servers and Web Applications with the intention of extracting sensitive data using the Cyber Kill Chain framework
• Penetration Test and Vulnerability Assessment
• Data collection, data analysis (logs, alerts, etc.) and reporting using the IBM Security QRadar SIEM