Positions, education, and fellowships. Full CV (PDF)
MIT, Laboratory for Nuclear Science, Cambridge, USA
Built Bayesian, multi-fidelity emulators for ab initio nuclear structure (VS-IMSRG) with full uncertainty quantification. Bayesian calibration and Shapley sensitivity of chiral EFT couplings. First laser-spectroscopy measurements of charge radii near the proton drip line. Large-scale calculations on HPC (SubMIT, ORNL); DAQ design and radioactive-molecule experiments with Harvard and Caltech.
VISA
Data and AI Teams. Published at ACM ICAIF and poster at’25,
CERN, Vanderbilt University, Geneva / Nashville
b-tagging and anomaly detection for the luminosity detectors; data pipelines for the BRIL system. Beyond-Standard-Model dark-matter studies at the LHC.
Guane Enterprises, Colombia
Joined as a machine learning engineer while an undergraduate and was promoted twice within two years, ultimately leading the data science team at Colombia’s most innovative startup of the period. Left the role to take up research with the CMS Collaboration.
University of British Columbia, Vancouver, Canada
Atomic Cluster Expansion methods and symmetry-conserving models, including representations for non-compact groups.
EIA University, Universidad de Medellín, Colombia
Neutrino oscillation phenomenology at DUNE and scalar dark-matter models. Geometric methods for particle reconstruction; physics-informed solvers for the Schrödinger equation.
Massachusetts Institute of Technology, Cambridge, USA
García Ruiz group, Laboratory for Nuclear Science
My research connects nuclear theory, machine learning, and precision experiments. I've developed Bayesian emulators for ab initio nuclear calculations, using them to constrain nuclear interactions and understand their uncertainties. More recently, I've worked on neural quantum states, molecular spectroscopy and control, and digital twins for experimental optimization. My work also involves laser spectroscopy, data acquisition, and high-performance scientific computing, in collaboration with experimental and theoretical groups across several institutions.
EIA University, Envigado, Colombia
GPA 4.9 / 5.0, top of class.
Universidad de Antioquia, Medellín, Colombia
Summit for AI Institutes Leadership (SAIL), San Diego, California, USA
Poster
MIT Generative AI Impact Consortium Symposium, Cambridge, Massachusetts, USA
Poster
CCMI — Cold and Controlled Molecules and Ions, Port Jefferson, New York, USA
Poster
IAIFI Summer Workshop, Cambridge, Massachusetts, USA
Talk Physics Driven Global Emulation of Nuclear Ab-initio Calculations
IAIFI PhD Summer School, UMass Boston, Boston, Massachusetts, USA
Talk Symbolic Regression · Tutorial lead
ECT* DTP/TALENT — Nuclear Theory for New Physics, Trento, Italy
Talk The moment for nuclear moments
Workshop on Experimental Design & Digital Twins, Universidad Nacional de Colombia, Medellín, Colombia
Talk Instructor
NPTwins — Digital Twins for Nuclear and Particle Physics, Trento, Italy
Talk Linking Electromagnetic Moments to Nuclear Interactions via Global Parametric ab-initio Emulation
INT 26-1 — Nuclear Hamiltonians for Advancing Nuclear Physics and Beyond, Seattle, Washington, USA
Talk Framing the problem differently: The perspective of Emulation
PAINT 2026 — Progress in Ab Initio Nuclear Theory, Vancouver, Canada
Attended
Quantum Science and Artificial Intelligence for Fundamental Physics, Medellín, Colombia
Talk Nuclear models beyond black boxes · Also co-chaired the student panel
NeurIPS Workshop on Generative AI in Finance, San Diego, California, USA
Talk Compliant Generative Diffusion for Finance
ACAT 2025, Hamburg, Germany
Talk BAyesian Neural Network: an Atomic and Nuclear Emulator (BANNANE)
IAIFI Summer Workshop, Cambridge, Massachusetts, USA
Talk Enforcing Hard Constraints in Generative Models for Mixed-Type Tabular Data
IAIFI Summer Workshop, Cambridge, Massachusetts, USA
Poster
Nuclear & Particle Physics Seminar, University of Vienna, Vienna, Austria
Talk Nuclear many-body from ab-initio interactions · Invited talk
LNS Physics Colloquium, MIT, Cambridge, Massachusetts, USA
Talk
Experimental Design, Sampling, and Optimization Strategies in Uncertainty Quantification, Chicago, Illinois, USA
Attended
PAINT 2025 — Progress in Ab Initio Nuclear Theory, Vancouver, Canada
Talk What is there inside a BANNANE: ML Tools for Global Emulation
MIT subMIT Computing Workshop, Cambridge, Massachusetts, USA
Talk Symbolic Learning Nuclear Relations
IAIFI Seminar, MIT, Cambridge, Massachusetts, USA
Talk Equivariant ML for Nuclear and Particle Physics
Bridging Talents & Opportunities Forum, Cambridge, Massachusetts, USA
Attended
APS Division of Nuclear Physics Fall Meeting, Boston, Massachusetts, USA
Talk Description of nuclear properties using Symbolic Machine Learning
AI+Science Summer School, Chicago, Illinois, USA
Poster
High Energy Seminar, University of Florida, Gainesville, Florida, USA
Talk Invited seminar
INT 24-1 — Fundamental Physics with Radioactive Molecules, Seattle, Washington, USA
Attended Program participant
PAINT 2024 — Progress in Ab Initio Nuclear Theory, Vancouver, Canada
Talk G-Equivariant Architectures for Many-Body Systems
AI4Phys — AI-driven discovery in physics and astrophysics, Kashiwa, Japan
Talk Symbolic Expansion: A Novel Approach to Symbolic Regression and Its Applications in Effective Physical Model Building
Molecular Machine Learning Conference (MoML), Cambridge, Massachusetts, USA
Attended
MIT–Amazon Science Hub Research Showcase, Cambridge, Massachusetts, USA
Attended
CMS BTV Workshop — To b or not to b, Brussels, Belgium
Talk G-Invariant GNNs for Jet Tagging
MOCa 2023 — Materia Oscura en Colombia, Medellín, Colombia
Talk Enhancing Parameter Exclusion and Extrapolation in Physics Applications with Advanced Machine Learning Techniques: A Phenomenological Perspective
ML4Jets 2022, Online · Rutgers University, New Jersey, USA
Talk Boost-Invariant Polynomials: an efficient and interpretable approach to jet tagging
ACAT 2022, Bari, Italy
Talk Boost-Invariant Polynomials: an efficient and interpretable approach to jet tagging
BOOST 2022, Hamburg, Germany
Talk Boost-Invariant Polynomials: an efficient and interpretable approach to jet tagging
RedCOLSI National and International Meeting (ENISI), Online
Talk Estudio de las condiciones necesarias para acotar el vacío por debajo en extensiones del Modelo Estándar con partículas escalares de materia oscura, usando el criterio de co-positividad
RedCOLSI Antioquia, Online
Talk Estudio de las condiciones necesarias para acotar el vacío por debajo en extensiones del Modelo Estándar con partículas escalares de materia oscura, usando el criterio de co-positividad