AI for outer space exploration · MIT AeroAstro (PhD student) · Former NASA JPL Visiting Student Researcher
I'm a PhD student at MIT AeroAstro, advised by Dr. Dava Newman and Dr. Daniel Varon. My interests sit at the intersection of artificial intelligence and space exploration, from remote sensing and Earth observation to autonomy for planetary exploration.
I've worked on projects spanning mineral classification from VNIR spectra at ESA European Astronaut Centre to physics-informed machine learning for lunar rover motion prediction at NASA JPL. Also at JPL I developed computer vision to detect internal waves in SWOT ocean imagery, work that earned a JPL Spotlight Technology Award. At Frontier Development Lab I contributed to Lunar-FM, the first AI foundation model for Moon exploration.
I'm an AI and Aerospace Engineer with an industry background in software and test engineering. I also hold a second Master's specialized in Artificial Intelligence from Mines Paris / PSL University. My research combines deep learning, computer vision, physics-informed machine learning and systems engineering to solve complex challenges in space science and exploration.
At MIT Responsive Environments (Dr. Joseph Paradiso), I developed AI architectures to correlate entangled signals captured by embedded smart fibers designed for the ISS, focusing on robust multi‑modal sensing, alignment, and interpretable inference under resource constraints.
MIT Media Lab, Responsive Environments, 2025.
Built a CV pipeline to detect and track internal waves (solitons) in SWOT imagery to support large‑scale ocean studies; recognized with a JPL Spotlight Technology Award.
American Geophysical Union (AGU), 2024 (oral) · AGU, 2024 (abstract)
Interoperable digital twins in Omniverse with FEM ground truth to train fast, physically consistent motion predictors for granular terrains and traverse planning.
IEEE Aerospace Conference, 2025 (two papers) · International Astronautical Congress (IAC), 2024 (interactive).
As part of the HWO AI working group, I contributed to the design of ML tools across two tracks: (1) enhancing mission design & operations and (2) models and tools for processing/analysis of observational data.
AGU, 2024 abstract (mission design & operations) · AGU, 2024 abstract (AI/ML models & tools)
Team project to learn EUV spectral irradiance proxies from SDO/AIA images when direct spectrograph coverage is unavailable.
American Geophysical Union (AGU), 2024 (abstract) · FDL (program)
2024: Oral presentation of “MEGS‑AI…” at AGU 2024, Washington, D.C.
Team effort to reconstruct time‑varying 3D density and temperature of the corona using a NeRF‑style approach.
American Geophysical Union (AGU), 2024 (abstract) · FDL (program)
Team project to unify GRAIL, LRO/LOLA, MiniRF, Clementine into a queryable foundation model for lunar science and operations.
ICLR 2026 Workshop FM4Science · arXiv:2607.22408 · LunarLab.ai
1D convolutional classifier for spectrometer time‑series to support astronaut sampling; improved accuracy and reduced variance vs. a prior CNN baseline. Report is internal/confidential.
European Astronaut Centre (ESA), 2022 (internal report).
Feasibility study to convert HLS into a long‑term lunar habitat, covering structural, life‑support, human‑factors and ops considerations.
Project webpage · IAC 2021–2023