Research

• Generative AI (GANs, diffusion, LLMs); trustworthy & robust AI; simulator-driven testing; oracle-free evaluation; clustering and explainability; safety-critical perception (space, autonomous driving); scientific & educational applications of generative models.

• European Space Agency (ESA) — Safety Analysis in Space Systems Lead Researcher, 2023–2025

Simulator-driven testing and retraining for planetary exploration perception without ground truth. GAN/diffusion-based input generation; evolutionary search; surrogate oracles (flip, uncertainty, clustering). Outcomes: TSE/TOSEM publications; MarsSim Simulator; ESA TIA deliverables.

• IEE Sensors — Safety Analysis in Automotive Systems Technical Contributor, 2021–2023

Root-cause analysis of DNN failures in ADAS perception. Input generation and explainability tooling to improve robustness and safety metrics; methodology transfer to industrial validation pipelines.

PhD (co-supervision, Uni.lu): 

• AI safety; generative models; simulation-based testing (2021–2023).

MSc Theses (selected):

• Image-to-Image Translation for Realism Enhancement: Pix2Pix/CycleGAN for simulator↔real bridging.

• Deep Embedded Clustering for Unsupervised Image Segmentation: representation learning for label-free segmentation.

• Simulation-driven DNN Testing: MarsSim + generative models for failure-inducing input generation.

• Uncertainty Estimation for Safety-Critical AI : Monte Carlo dropout for robustness assessment.

• Program Committe Member: GECCO conference 2024-2025.

• Reviewer: IEEE TSE, TOSEM, EMSE, GECCO, etc.

• Program Committe Member: ISSRE conference 2022.

• Organizer: MAGI Workshop Scientific computing, cloud, big data, Villetaneuse-France, October 2019