Luisa Orozco
PhD
Data Scientist and Software Engineer
About Me
I am a data scientist and software engineer specialized in applying machine learning to material science. My background is in mechanics of granular matter complemented by experience on improving software practices in research. At TNO - GDN in Utrecht I currently apply many data science, AI and machine learning tools to problems around geo data management of the underground of the Netherlands.
I believe in Open Science and try to advocate for it by sharing the source code of my projects and contributing to open-source libraries and projects.
Keywords about my work:- ML for material science
- Granular matter
- AI for Geoscience
- MLOps
- Geotechnics
- FOSS
Experience
I am part of the GeoData & Information Management (GDIM) department and as a member of the Science team, I contribute on projects necessitating Machine learning, AI, or big data processing tools crafted to challenging problems.
As an RSE at NLeSC I collaborated with researchers to develop software tools that enhance their research. I am part of the Natural Sciences section and team Nabla \( \nabla \) working on projects involving Machine Learning and materials science.
Projects I have contributed to:
- GrainLearning: Neural networks as surrogates for simulation parameters inference.
- OnSET: Improving software practices for an efficient and robust open-source modelling platform for research on Energy Transition applications.
- DeepMolGen: Exploring generative AI tools for molecule generation with optimal properties.
- DEEPDIP: Enhance ML aided models with differentiable programming in Julia.
At GRASP (Group of Research and Application in Statistical Physics), I was part of AEROPERF a multidisciplinary team focusing on the development and optimization of inhalation powders. These powders were engineered for effective pulmonary drug delivery, ensuring deep and homogeneous distribution in the lungs. My role involved the physical characterization of these spray-dried powders, specifically examining their agglomeration to improve flowability. Additionally, I conducted experiments and simulations on air fluidized beds, contributing to a deeper understanding of physical mechanisms crucial in various industrial applications.
My activities at the Laboratory of Mechanics and Civil Engineering (LMGC) included:
- In collaboration with INRAE, SupAgro supervise a Master student on experiments on grinding of a granular material.
- Research on effects of particle shape on the flow and breakage of granular materials.
- Teaching tasks included laboratory of materials and civil engineering.
Modelling and characterizing granular flows in rotating drums and to elucidate the mechanisms of particle grinding across scales. By means of particle dynamics simulations and a breakage model involving the tessellation of particles and a dynamic debonding criterion based on fracture mechanics, I investigated single-particle impacts, the properties of granular flows in the cascading regime in a rotating drum as a function of system parameters, the effect of the size and amount of grinding media on the grinding process, and the evolution of particle breakage in flows of breakable particles.
Education
2016 - 2019
PhD in Mechanics and Civil Engineering
CEA, LMGC, Université de Montpellier
By means of numerical simulation of breakage of granular materials using the Discrete Element Method (DEM), I investigated single-particle impacts, the properties of granular flows in the cascading regime in a rotating drum as a function of system parameters, the effect of the size and amount of grinding media on the grinding process, and the evolution of particle breakage in flows of breakable particles.
2015 - 2016
Master of Science in Civil Engineering - Geotechnics
Universidad de los Andes
I performed experimental studies on the effects of hydrophobic particles on water migration in fine graded soils. This project was in the framework of a collaboration between the Universidad de los Andes and the Government of Vichada, Colombia. The goal was to improve the soil properties of the region to enable the construction of roads and other infrastructure.
Projects
Scientific Machine learning
Navier-Stokes
Julia
CWI
DEEPDIP
In collaboration with the Scientific Computing group at CWI, we investigate the use of neural closure models for Navier-Stokes equations.
Generative AI
Chemistry
UvA
DeepMolGen
In collaboration with Ensing research group at HIMS - UVA, we investigate the use of Diffusion models for generation of crystal structures.
FOSS
CMake
TUDelft
OnSET
Open-darts is a modelling framework to simulate multi-physics processes in geo-engineering applications such as: geothermal, CO2 sequestration and hydrogen storage.
PhD thesis
DEM
Impact Particle breakage in DEM
Impact of particles on a rigid plane. The simulations are performed using the Contact Dynamics (CD) algorithm with a Bonded-Cell Method (BCM) for modelling particle breakage.
Particle engineering
Powder characterization
Engineering particles to enhance pulmonary drug delivery
In collaboration with pharmacists and chemical engineers we investigated the effect of carbohydrate excipient on the aerosolization of spray-dried powders.
PhD thesis
DEM
Particle breakage inside a ball mill
By means of DEM simulations of the continuous breakage of multiple particles inside a rotating cylinder, we analyzed the effect of several choices such as the rotation speed, size of grinding media, filling degree. We proposed a scaling-law for the grinding process based on our studies.