Joseph Gallego

joseph-gallego

Profile

Assistant Teaching Professor at Drexel. Postdoctoral researcher at Trillium, funded by NASA and the European Space Agency. PhD, Master's, and Bachelor's in computer science and electrical engineering with a focus on machine learning. Participant and winner of competitive programming at the university and high school.

14 years of experience doing research at the graduate level and 22 years from his first line of code. Published 39 research articles, including 6 journal articles in Q1 journals, and received a meritorious distinction for his doctorate. Served as a reviewer for 8 journals/conferences, spanning Q1 journals and A1 conferences. Holds a teaching assistant professor position at Drexel University, an R1 US institution. His work has been recognized with a Best Paper Award at a NeurIPS workshop. 10 years of experience teaching university courses, spanning undergraduate and graduate courses. Additionally, he spent one year and two months at FDL doing research as a postdoctoral fellow and then 6 months as a faculty leader on NASA and European Space Agency projects. He has coauthored articles with researchers from Oxford and MIT. Advised and guided eight researchers, 1 at the postdoc level, 1 at the Ph.D. level, 4 at the master's level, and 2 at the undergraduate level. A rate of 5/5 according to Rate my Professors, 4.5/5 according to the last survey grade received at Drexel University, and 4.6/5 at the National University of Colombia. Served as a jury in programming, machine learning, and entrepreneurship contests. Ex-software engineering manager of Munkys Group INC. It is a tech company. Developed applications where there is a broad component of computer science, computer and electrical engineering, cyber-security, statistics, machine learning, and industrial engineering, using principles of Lean Manufacturing and statistics for software development.

I am passionate about programming and the use of free technologies for the rapid development of web and mobile applications. Leader, with aptitudes and skills to work in a team, take responsibility, create a commitment, and of course, always be willing to change. I am interested in generating new knowledge. I teach both programming and artificial intelligence, one of the activities that I enjoy the most.

Education

Frontier Development Lab ‑ NASA ‑ ESA

Postdoc machine learning research US - Europe June. 2023 - Aug. 2024

Development of Worldwide machine learning foundational response models

and build machine learning models for forest height estimation.

Development of models for height tree estimation using tomographic

synthetic aperture radar images.

National University of Colombia Bogotá

PhD with meritorious distinction in systems and computing engineering Colombia Jul. 2019 - Jun. 2023

Grade point average 3.9 out of 4.0.
Teaching assistant scholarship.

Deep.AI

Specialization in deep learning Bogotá, Colombia Jan. 2020 - Dec. 2020

6 Courses approved in deep learning.
Coursera platform.

National University of Colombia Medellín

Master's Degree in Systems Engineering and Computer Science Colombia Jan. 2014 - Dec. 2016

Grade point average 3.88 out of 4.0.
Teaching assistant scholarship.
Young Researcher.

National University of Colombia Medellín

Systems and computer engineering National University of Colombia Bogotá, Colombia Jan. 2008 - Dec. 2013

Grade point average 3.74 out of 4.0.
Winner of "mejores saber pro".
Part of the Dean List every semester.
Third place in graduation in systems and computer engineering.

National University of Colombia Bogotá

Industrial engineering, Colombia Jan. 2008 - Dec. 2013

Grade point average 3.66 out of 4.0.
Winner of "mejores saber pro".
Part of the Dean List every semester.
First place in industrial engineering graduation.

Technician Center Technician Institute of La Salle Bogotá

Intermediate Education - Systems , Colombia Jan. 2003 - Dec. 2007

Best student in school.
Participant and winner of systems and computer olympiads.

Papers

CORONA-FIELDS: Classifying Solar Wind Structures from In-Situ Plasma Parameters with Masked Autoencoders and Neural Fields

AAAI 2026

Jinsu Hong, Daniela Martin, Connor O'Brien, Valmir Moraes, Jasmine Kobayashi, Evangelia Samara, Joseph Gallego

In this work, we leverage a foundation model for solar physics trained on the Solar Dynamics Observatory images, using its Masked Autoencoder variant to produce embeddings suitable for solar wind structure classification. Leveraging the produced embeddings, we perform fine-tuning by incorporating a neural field--based architecture that encodes spacecraft position and solar magnetic connectivity in a high-frequency space.

Kernel Density Matrices

Quantum Machine Intelligence 2025

Fabio A González, Raul Ramos,Joseph A Gallego

This paper presents a novel approach to probabilistic deep learning (PDL), quantum kernel mixtures, derived from the mathematical formalism of quantum density matrices, which provides a simpler yet effective mechanism for representing joint probability distributions of both continuous and discrete random variables. - Click Here - Research Gate

Uncovering Solar Wind Phenomena with iSAX, HDBScan, Human-in-the-loop and PSP Observations

NeurIPS 2025 Workshop AI4Science 2025

Valmir P Moraes Filho, Daniela Martin, Jasmine R. Kobayashi, Connor O'Brien, Jinsu Hong, Evangelia Samara, Joseph Gallego

We introduce a pipeline that combines symbolic compression and indexing, unsupervised density-based clustering, and human-in-the-loop validation. Applied to 2018-2024 PSP measurements, this framework efficiently processes over 150 GB of magnetic and plasma data, reducing computational cost while preserving physical interpretability. The method successfully recovers known solar wind structures, identifies uncatalogued CMEs and transient events, and demonstrates robustness across multiple time scales.

CIPHER: Scalable Time Series Analysis for Physical Sciences with Application to Solar Wind Phenomena

NeurIPS 2025 Workshop AI4Science 2025

Jasmine R. Kobayashi, Daniela Martin, Valmir P Moraes Filho, Connor O'Brien, Jinsu Hong, Hala Lamdouar, Sairam Sundaresan, Anna Jungbluth, Sudeshna Boro Saikia, Andrés Muñoz-Jaramillo, Evangelia Samara, Joseph Gallego

We present the Clustering and Indexation Pipeline with Human Evaluation for Recognition (CIPHER), a framework designed to accelerate large-scale labeling of complex time series in physics. CIPHER integrates indexable Symbolic Aggregate approXimation (iSAX) for interpretable compression and indexing, density-based clustering (HDBSCAN) to group recurring phenomena, and a human-in-the-loop step for efficient expert validation. R

Scalable Machine Learning Analysis of Parker Solar Probe Solar Wind Data

NeurIPS 2025 Workshop AI4Science 2025

Daniela Martin, Connor O'Brien, Valmir P Moraes Filho, Jinsu Hong, Jasmine R. Kobayashi, Evangelia Samara, Joseph Gallego

Our framework leverages Dask for large-scale statistical computations and KDM to estimate univariate and bivariate distributions of key solar wind parameters, including solar wind speed, proton density, and proton thermal speed, as well as anomaly thresholds for each parameter. We reveal characteristic trends in the inner heliosphere, including increasing solar wind speed with distance from the Sun, decreasing proton density, and the inverse relationship between speed and density. Solar wind structures play a critical role in enhancing and mediating extreme space weather phenomena and can trigger geomagnetic storms; our analyses provide quantitative insights into these processes.

Sea Ice Drift Estimation in the Arctic: Deep Learning Optical Flow on RADARSAT-2

NeurIPS 2025 Workshop AI4Science 2025

Daniela Martin, Joseph Gallego Towards Reliable

We present the first large-scale benchmark of 48 deep learning optical flow models on RADARSAT-2 ScanSAR data, evaluated with endpoint error (EPE) and Fl-all metrics against GNSS-tracked buoys. Several models achieve sub-kilometer accuracy (EPE 6-8 pixels, 300-400m), capturing coherent regional drift patterns and demonstrating that deep learning computer vision models can be effectively transferred to polar remote sensing.

INQMAD: Incremental Streaming Anomaly Detection with Density Matrices, Quantum Measurement and Density Estimation

Neural Computing and Applications 2024

Joseph A Gallego, Oscar A Bustos, Fabio A González

In our research, we propose a novel method for streaming anomaly detection based on incremental density estimation. The method uses random Fourier features and incorporates the principles of quantum measurements and density matrices. We present the definition of the method and a variant of it, that can be resembled as an exponential moving average density and a simple moving average, respectively. The method is designed to handle potentially infinite data streams and has a constant update complexity of $O(1)$.

M3LEO: A Multi-Modal, Multi-Label Earth Observation Dataset Integrating Interferometric SAR and RGB Data

Neurips 2025

Matthew Allen, Francisco Dorr, Joseph Alejandro Gallego Mejia, Laura Martínez-Ferrer, Anna Jungbluth, Freddie Kalaitzis, Raúl Ramos-Pollán

We introduce M3LEO, a multi-modal, multi-label EO dataset that includes polarimetric, interferometric, and coherence SAR data derived from Sentinel-1, alongside Sentinel-2 RGB imagery and a suite of labelled tasks for model evaluation. M3LEO spans 17.5TB and contains approximately 10M data chips across six geographic regions. The dataset is complemented by a flexible PyTorch Lightning framework, with configuration management using Hydra. We provide tools to process any dataset available on popular platforms such as Google Earth Engine for integration.

Quantum density estimation with density matrices: Application to quantum anomaly detection

Physical Review 2024

Diego H. Useche, Oscar A. Bustos-Brinez, Joseph A. Gallego, Fabio A. González

In this article, we present a quantum-classical density-matrix density estimation model, called Q-DEMDE, based on the expected values of density matrices and a quantum embedding called quantum Fourier features. The method uses quantum hardware to build probability distributions of training data via mixed quantum states.

Tree Height Estimation using Machine Learning and 3D Tomographic SAR-a case study in Northern Europe

AGU 2024

Joseph A Gallego-Mejia, Grace Colverd, Jumpei Takami, Laura Schade, Karol Bot

In this study, we use deep learning to estimate forest canopy height directly from 2D Single Look Complex (SLC) images, a derivative of SAR. Our method bypasses traditional tomographic processing, potentially reducing latency from SAR capture to end product. We quantify the impact of varying numbers of SLC images on height estimation accuracy, aiming to inform future satellite operations and optimize data collection strategies. Our results indicate a potential reduction in MAE in increasing the number of SAR inputs from 3 to 7 of 17%.

Tomographic SAR Reconstruction for Forest Height Estimation Neurips

Machine Learning for Science 2024

Grace Colverd, Jumpei Takami, Laura Schade, Karol Bot, Joseph A Gallego-Mejia

Tree Species Classification using Machine Learning and 3D Tomographic SAR - a case study in Northern Europe

AGU 2024

Joseph A Gallego-Mejia, Grace Colverd, Jumpei Takami, Laura Schade, Karol Bot

In this study, we employed TomoSense, a 3D tomographic dataset, which utilizes a stack of single-look complex (SLC) images, a byproduct of SAR, captured at different incidence angles to generate a three-dimensional representation of the terrain. Our research focuses on evaluating multiple tabular machine-learning models using the height information derived from the tomographic image intensities to classify eight distinct tree species. The SLC data and tomographic imagery were analyzed across different polarimetric configurations and geosplit configurations.

Tree Species Classification using Machine Learning and 3D Tomographic SAR - a case study in Northern Europe

NeurIPS Climate Change 2024

Grace Colverd, Jumpei Takami, Laura Schade, Karol Bot, Joseph A Gallego-Mejia

3D-SAR Tomography and Machine Learning for High-Resolution Tree Height Estimation

Arxiv 2024

Grace Colverd, Jumpei Takami, Laura Schade, Karol Bot, Joseph A Gallego-Mejia

This study applies machine learning to extract forest height data from two SAR products: Single Look Complex (SLC) images and tomographic cubes, in preparation for the ESA Biomass Satellite mission. We use the TomoSense dataset, containing SAR and LiDAR data from Germany's Eifel National Park, to develop and evaluate height estimation models. Our approach includes classical methods, deep learning with a 3D U-Net, and Bayesian-optimized techniques.

Towards Foundation Models for Earth Observation; Benchmarking Datasets and Performance on Diverse Downstream Tasks

AGU 2024

Matt Allen, Joseph A. Gallego-Mejia, Anna Jungbluth, Laura Martínez-Ferrer, Francisco Dorr, Freddie Kalaitzis, Raúl Ramos-Pollán

Latent Anomaly Detection Through Density Matrices

ArXiv 2024

Joseph A Gallego, Oscar A Bustos, Fabio A González

This paper introduces a novel anomaly detection framework that combines the robust statistical principles of density-estimation-based anomaly detection methods with the representation-learning capabilities of deep learning models. The method originated from this framework is presented in two different versions: a shallow approach employing a density-estimation model based on adaptive Fourier features and density matrices, and a deep approach that integrates an autoencoder to learn a low-dimensional representation of the data.

Exploring DINO: Emergent Properties and Limitations for Synthetic Aperture Radar Imagery

Neurips 2023

Joseph A. Gallego-Mejia, Anna Jungbluth, Laura Martínez-Ferrer, Matt Allen, Francisco Dorr, Freddie Kalaitzis, Raúl Ramos-Pollán

Self-supervised learning (SSL) models have recently demonstrated remarkable performance across various tasks, including image segmentation. This study delves into the emergent characteristics of the Self-Distillation with No Labels (DINO) algorithm and its application to Synthetic Aperture Radar (SAR) imagery. - Click Here - Research Gate

Fewshot learning on global multimodal embeddings for earth observation tasks

Neurips 2023

Matt Allen, Francisco Dorr, Joseph A Gallego-Mejia, Laura Martínez-Ferrer, Anna Jungbluth, Freddie Kalaitzis, Raúl Ramos-Pollán

In this work we pretrain a CLIP/ViT based model using three different modalities of satellite imagery across five AOIs covering over 10% of the earth total landmass, namely Sentinel 2 RGB optical imagery, Sentinel 1 SAR amplitude and Sentinel 1 SAR interferometric coherence. Click here - Research Gate

Exploring Generalisability of Self-Distillation with No Labels for SAR-Based Vegetation Prediction

Neurips 2023

Laura Martínez-Ferrer, Anna Jungbluth, Joseph A Gallego-Mejia, Matt Allen, Francisco Dorr, Freddie Kalaitzis, Raúl Ramos-Pollán

In this work we pre-train a DINO-ViT based model using two Synthetic Aperture Radar datasets (S1GRD or GSSIC) across three regions (China, Conus, Europe). We fine-tune the models on smaller labeled datasets to predict vegetation percentage, and empirically study the connection between the embedding space of the models and their ability to generalize across diverse geographic regions and to unseen data. - Click Here - Research Gate

Large Scale Masked Autoencoding for Reducing Label Requirements on SAR Data

Neurips - Climate Change - Best Paper Award 2023

Matt Allen, Francisco Dorr, Joseph A Gallego-Mejia, Laura Martínez-Ferrer, Anna Jungbluth, Freddie Kalaitzis, Raúl Ramos-Pollán

In this work, we apply a self-supervised pretraining scheme, masked autoencoding, to SAR amplitude data covering 8.7% of the Earth's land surface area, and tune the pretrained weights on two downstream tasks crucial to monitoring climate change - vegetation cover prediction and land cover classification. We show that the use of this pretraining scheme reduces labelling requirements for the downstream tasks by more than an order of magnitude, and that this pretraining generalises geographically, with the performance gain increasing when tuned downstream on regions outside the pretraining set. - Click Here - Research Gate

DEMANDE: Density Matrix Neural Density Estimation

IEEE ACCESS 2023

Joseph A Gallego, Fabio A González

This paper presents a novel method for neural density estimation based on density matrices and adaptive Fourier features. Density matrices are commonly used in quantum mechanics to represent the quantum state of a physical system. - Click Here - Research Gate

Learning with Density Matrices and Random Features

Quantum Machine Intelligence 2022

Fabio A González, Joseph A Gallego, Santiago Toledo-Cortés, Vladimir Vargas-Calderón

This paper explores how density matrices can be used as a building block to build machine learning models exploiting their ability to straightforwardly combine linear algebra and probability. Quantum Machine Intelligence - Click Here - Research Gate

Computing expectation values of density matrices for quantum anomaly detection

Arxiv 2022

Diego H. Useche, Oscar A. Bustos-Brinez, Joseph A. Gallego, Fabio A. González

Fast Kernel Density Estimation with Density Matrices and Random Fourier Features

Ibero-American Conference on Artificial Intelligence 2022

Joseph A Gallego, Juan F Osorio, Fabio A González

In this paper, we systematically evaluate the novel DMKDE algorithm and compare it with other state-of-the-art fast procedures for approximating the kernel density estimation method on different synthetic data sets. Ibero-American Conference on Artificial Intelligence - Click Here - Research Gate

Anomaly Detection through Density Matrices and Kernel Density Estimation (AD-DMKDE)

LatinX

Workshop in Neurips 2022 2022

Oscar A Bustos, Joseph A Gallego, Fabio A González

This paper presents a novel density estimation method for anomaly detection using density matrices (a powerful mathematical formalism from quantum mechanics) and Fourier features. Neurips 2022 LatinX Workshop - Click Here - Research Gate

InQMeasurement: Incremental Quantum Measurement Anomaly Detection

ICDM 2022

Joseph A Gallego, Oscar A Bustos, Fabio A González

We present a new incremental quantum measurement anomaly detection method based on Fourier features and density matrices. The IEEE International Conference on Data Mining (ICDM) - Click Here - Research Gate

LEAN-DMKDE: Quantum Latent Density Estimation for Anomaly Detection

AAAI 2023

Joseph A Gallego, Oscar A Bustos, Fabio A González

The method combines an autoencoder, for learning a low-dimensional representation of the data, with a density-estimation model based on random Fourier features and density matrices in an end-to-end architecture that can be trained using gradient-based optimization techniques. Arxiv - Click Here- Research Gate

Risk Automatic Prediction for Social Economy Companies using Camels

WEA 2022

Joseph A Gallego, Daniela Martin V, Fabio A González

We proposed a prediction model based on a machine learning approach. The method was trained with the random forest algorithm with historical data provided by each SEE. Three consecutive periods of data were concatenated. WEA - Click Here - Research Gate

MLOps (Machine Learning Dev Ops)

ACIS 2022

Joseph A Gallego, Fabio A González

Data analytics and machine learning are changing every aspect of our lives today. Activities that were once performed entirely by humans such as disease detection, object detection, speech recognition and synthesis, among others, are now being automated using machine learning algorithms. https://sistemas.acis.org.co/index.php/sistemas/article/view/205

"Machine Learning Operations and its implementation challenges in Colombia"

ACIS 2022

Joseph A Gallego, Fabio A González National Survey

Robust kernels for robust location estimation

Neurocomputings 429, 174-186 2021

Joseph A Gallego, Fabio A González, Olfa Nasraoui

This paper shows that least-square estimation (mean calculation) in a reproducing kernel Hilbert space (RKHS) F corresponds to different M-estimators in the original space depending on the kernel function associated with F. Neurocomputings - Click Here - Research Gate

Robust Estimation in Reproducing Kernel Hilbert Space

Neurips 2023 LatinX Workshop 2019

Joseph A Gallego, Fabio A González, Olfa Nasraoui

Our work shows that estimating the mean in a feature space induced by certain kinds of kernels is the same as doing a robust mean estimation using an M-estimator in the original problem space - Click Here - Research Gate

Work Experience

Trillium - Frontier development lab - NASA US

Faculty: Machine Learning Leader 04/2025 - 12/2025

Create model for solar wind
Use of model versioning with Hydra
Using Google Cloud Platform for training using GPUs
Using Nvidia's DGX for multi-gpu and distributed multi-gpu training
Reading and implementing state-of-the-art models
Daily work progress meetings
Use of git for code versioning
Use of best development practices

Contact: Anne Jungbluth - Email: anne@trillium.tech

Drexel University Philadelphia

Assistant Teaching Professor, US 09/2024

Teaching CS265: Advanced programming course and techniques
Teaching CS270: Mathematical Foundations
Teaching CS260: Data Structures and Algorithms
Teaching CS521: Introduction to algorithms
Teaching CS660: Data Science at Scale

Drexel University Philadelphia

Adjunct Professor,

Philadelphia, US 04/2024 - 09/2024

Teaching CS265: Advanced programming course and techniques
Teaching CS502: Data structures and algorithms
Teaching CS521: Introduction to algorithms

Munkys Group Inc

Software and Machine Learning Engineer Manager

Wilmington, US 01/2012 - 07/2024

Create a full machine learning pipeline using PyTorch and Pytorch Lightning for Generative AI using stable diffusion models and large language models for quality texture generated images which are used by 10k users.
Create a Spring Boot application for manufacturing enterprises supporting 10k mobile devices users with 10M transactions each hour, using PostgreSQL, RabbitMQ and redis.
Create 10 android applications using Java and Kotlin that supports 10k users. The applications has several usecases such as maintenance, quality, production, messages, payroll, biometric.
Create a triplet network to improve the human recognition of the employees, improving the accuracy by 20% and reducing the inference time by 15%
Create a hybrid infrastructure using docker, docker compose, and kubernetes to support 10k users using a hybrid infrastructure. This achived a cost reduction of 60% compared to the traditional virtual machine cloud setting.
Lead a team of 8 developers and designers to create 10 applications using Scrum, with dealy meetings. Supporting, encourage, and solving their problems.
Creator and developer of production software with mobile devices and IOT.
Has designed databases, both SQL (Postgresql) and No-SQL (MongoDb
Has led the programming group using SCRUM methodology with 15-day sprints.
Has created machine learning models for classification, regression, anomaly detection tasks, among others.
Has created machine learning models for clustering tasks.
Has created machine learning pipelines using Tensorflow and Pytorch.
Has used MLflow and Weights&Biased for logging the models.
Has used Hydra to make the configuration of the models.
Has developed Front-End applications using technologies such as CSS, HTML, HTML5, Javascript, AngularJs, Angular7 and React.
Has developed Backend applications using technologies such as Java, Spring Boot, Jhipster, Flask among others.
Has developed mobile applications on native Android with Java and Kotlin.
Has used communication technologies between microservices with asynchronous communication with Apache Kafka.
Has implemented databases. Performed entity-relationship model generation for several of our products.
Has performed analysis of quantitative data generated from the Sammu application (use of tools such as linear regression, logistic regression, multifactor analysis, clustering, graphical report generation, classification).
Participated in the construction, elaboration and subsequent analysis of Sammu customer satisfaction surveys.
Processing of the information collected and generated by the Sammu application.
Used different descriptive statistics tools to generate graphical reports for our clients.
Use of computational learning tools to extract information from the data generated by Sammu customers.
Generation of technical documents used by our developers.
In charge of leading and supervising the weekly meetings held by the development team, where documents are generated.

Contact: Daniela Martin

Trillium - Frontier development

Postdoc: Machine Learning Researcher

lab - ESA - NASA US - Europe 06/2023 - 07/2024

Create machine learning pipeline using self-supervised learning on satellite images and improve the state of the art methods on 5% of accuracy
Collaborate with 7 machine learning researches to create new self Clip supervised machine learning methods to improve by 10 percent state of the art methods in vegetation and biomass estimation
Create a mask autoencoder algorithm to improve land cover classification using self-supervised learning methodology and improve by 7% the state-of-the-art-methods
Present the results to a large audience of experts in 10 minutes showing the principal results obtained by the team
Create a PyTorch lightning dataloader and dataset to ingress a variety of satellite images with different modalities using Geopandas, XArray and Dask and improve the reading time by 50% using Blosc compression.
Use of model versioning with Hydra
Using Google Cloud Platform for training using GPUs
Using Nvidia's DGX for multi-gpu and distributed multi-gpu training
Reading and implementing state-of-the-art models
Daily work progress meetings
Use of git for code versioning
Use of best development practices

Contact: Anne Jungbluth - Email: anne@trillium.tech

National University of Colombia Bogotá, Colombia

Computer Programming

08/2022 - 07/2023

Preparation of master classes, tests and activities.
Evaluation and grading of midterm exams.
Contact: Luis Fernando Niño - Phone: +57 300 269 3300

Center of the criminal investigation department and interpol Comware

Knowledge transfer lecturer in the project development of the predictive model of crime for the criminal analysis

Bogotá, Colombia 12/2021 - 01/2022

Pedagogical design, construction of educational content, planning and execution of the knowledge transfer program. execution of the knowledge transfer program
Pedagogical design, construction of educational content, planning and execution of the sessions and workshops of the machine learning module in the knowledge transfer program.
Elaboration of the instructional scripts and approach of the learning activities based on the competency map and what is described in the syllabus.
Design of the activities to be developed in a synchronous distance, active-participative and demonstrative-explanatory way, by means of strategies of adoption and use of emerging technologies that incorporate methodologies of learning-by-doing and application of knowledge to problems and cases of the real environment, making use of the concepts and best practices acquired in the development of these experiences.
Ensure that the knowledge transfer program is developed on the basis of meaningful learning experiences for participants, integrating at least, the following activities: - Guided workshops and discussion groups. - Virtual content, video tutorials and learning capsules. - Simulation and evaluation of real cases and contextual problems.
Conducting the keynote lectures for the four sessions that make up the machine learning module.

Contact: Comware - Phone: +57 1 638 21 00 :::

National University of Colombia

Teaching support for "Machine Learning And Big Data" diploma course

Bogotá, Colombia 08/2019 - 2024

Support to the teacher in teaching activities.
Grading of assignments, assignments and projects.
Development and remote face-to-face teaching.

Contact: Fabio Gonzalez - Phone: +57 300 2693621

Etraining

Machine Learning module leader and expert in charge of the sessions

Bogotá, Colombia 09/2021 - 10/2021, 05/2022 - 10/2022

Pedagogical design, construction of educational content, planning and execution of the knowledge transfer program.
Perform the pedagogical design, content construction, planning and execution of the sessions and workshops of the machine learning module in the program for knowledge transfer in BLOCKCHAIN AND DATA ANALYTICS FOR DIGITAL INDUSTRY.
Development of the instructional scripts and approach of the learning activities based on the competency map and what is described in the syllabus.
Design of the activities to be developed in a synchronous distance, active-participative and demonstrative-explanatory manner, through strategies of adoption and use of emerging technologies that incorporate learning-by-doing methodologies and application of knowledge to problems and cases of the real environment, making use of the concepts and best practices acquired in the development of these experiences.
Ensure that the knowledge transfer program is developed on the basis of meaningful learning experiences for participants, integrating at least, the following activities: - Guided workshops and discussion groups - Virtual content, video tutorials and learning capsules - Simulation and evaluation of real cases and contextual problems
Support in the development of the sessions for the execution of the knowledge transfer to the 220 beneficiary companies through the ENSÉÑAME platform provided by Etraining, in the topics under its responsibility. Consequently: - Conducting the master lectures for the four sessions that make up the machine learning module - Participating in the development of information collection instruments to beneficiary companies - Delivering inputs for reports
Delivery of inputs for the diagnostic reports prepared by Etraining on the beneficiary companies of the program in the topics under its responsibility according to the guidelines given by the person in charge delegated by Etraining for the preparation of the final report, as follows: - Participation in the structuring of instruments to identify opportunities for the application of machine learning in the value chain of each company - Participation in the structuring of instruments to identify opportunities for the application of machine learning in the business model - Review and analysis of information presented by the companies with initiatives for application on the value chain and business model.

Contact: Alexandra Hernández - Phone: +57 310 6899 9867

National University of Colombia

Data Structures Teacher

Bogotá, Colombia 08/2020 - 06/2022

Preparation of master classes, tests and activities.
Evaluation and grading of midterm exams.

Contact: Luis Fernando Niño - Phone: +57 300 269 3300

National University of Colombia

Creation of material for Agile Methodologies and Machine Learning course

Bogotá, Colombia 08/2021 - 12/2021

Development of interactive material, presentations, workshops, projects,

among others.

Contact: Fabio Gonzalez - Phone: +57 300 2693621

National University of Colombia - Colombian Technology Minister

Computer programming and software development lecturer -

Bogotá, Colombia 08/2021 - 12/2021

To provide professional services as a specific trainer cycle 1 and 2 in the framework of the inter-administrative agreement of the inter-administrative agreement No 782 of 2021, teaching classes in the languages Python and Java. Python and Java
Provide professional services as facilitator 1 level - cycle ii and iii in the framework of the project "contract for the provision of professional services as facilitator 1 level - cycle ii and iii in the framework of the project of the project "service provision contract no 0246 of 2020 subscribed between the Tecnalia Colombia Foundation and the national university of colombia".

Contact: Juan Carlos Torres Phone: +57 300 372 3702

General Comptroller's Office of the Republic

Natural Language Processing Lecturer National University of Colombia -

Bogotá, Colombia 08/2021 - 12/2021

To provide professional services as a lecturer within the framework of the project "inter-administrative contract 374 of 2020 subscribed between the comptroller general's office de la republica y la universidad nacional de colombia"
OSE No. 2409 de 2020 Empresa QUIPU: 2018 - 01/12/2020 a 10/12/2020

Contact: Felipe Restrepo - Phone: +57 304 5762504

University of Colombia - Supersolidaria Bogotá

Development of a machine learning system for Supersolidaria National

Colombia 08/2021 - 12/2021

Professional services to support machine learning analytics and solution development of the development of solutions of the Supersolidaria's Think Tank in the framework of the the framework of the project called "Structuring of a thinking center in solidarity economics in solidarity economy"

Contact: Hernan Ceballo - Phone: +57 312 5035901

National University of Colombia

Programming fundamentals lecturer

Medellín, Colombia 07/2015 - 05/2016

Preparation of master classes, quizzes and activities.
Evaluation and grading of midterms.

National University of Colombia

Young Researcher Minciencias -

Bogotá, Colombia 01/2014 - 12/2015

Development of an artificial intelligence system for video

Contact: Fabio Augusto González Phone: +57 300 269 36 21

Munkys SAS

Administrative and Financial General Manager

Bogotá, Colombia 01/2007 - 12/2011

Contact: Ilda María Mejía Duque Phone: +57 320 451 92 28

National University of Colombia

Office Assistant

Bogotá, Colombia 01/2010 - 06/2010

Computer maintenance, software installation, software update.

Software and Datasets

CORONA-FIELDS CORONA-Fields: Leveraging Foundation Models for Classification of Solar Wind Phenomena US 2025

Github: https://github.com/spaceml-org/CORONA-FIELDS

CIPHER CIPHER: Scalable Time Series Analysis for Physical Sciences with Application to Solar Wind PhenomenaC US 2025

Github: https://github.com/spaceml-org/CIPHER

PSP-KDM Scalable Machine Learning Analysis of Parker Solar Probe Solar Wind Data US 2025

Github: https://github.com/spaceml-org/PSP-KDM

M3LEO Software Backbone: M3LEO: A Multi-Modal Multi-Label Earth Observation Dataset US 2024

Github: https://github.com/spaceml-org/M3LEO

M3LEO: A Multi-Modal Multi-Label Earth Observation Dataset Dataset US 2024

Huggingface: https://huggingface.co/M3LEO

Demande Dataset Dataset US 2023

Zenodo: https://zenodo.org/records/7822851

Learning with Density Matrices and Random Fourier Features Incremental Anomaly Detection using Quantum Measurements US 2022

Github: https://github.com/Joaggi/Incremental-Anomaly-Detection-using-Quantum-Measurements

Learning with Density Matrices and Random Fourier Features LEAN-DMKDE: Quantum Latent Density Estimation for Anomaly Detection US 2022

Github: https://github.com/Joaggi/qaddemadac

Learning with Density Matrices and Random Fourier Features Anomaly Detection with Density Matrix Kernel Density Estimation US 2022

Github: https://github.com/Joaggi/anomaly-detection-density-matrix-kernel-density-estimation

Learning with Density Matrices and Random Fourier Features Fast Kernel Density Estimation US 2021

Github: https://github.com/Joaggi/Fast-Kernel-Density-Estimation-with-Density-Matrices-and-Random-Fourier-Features

Learning with Density Matrices and Random Fourier Features Quantum Measurement Classification US 2021

Github: https://github.com/fagonzalezo/qmc

Robust Kernels Robust Estimation in Reproducing Kernel Hilbert Space Bogota, Col 2019

Github: https://github.com/Joaggi/Robust-kernels-for-robust-location-estimation

Prototype of a Distributed Language for Agents Undergraduate Thesis Bogota, Col 2013

Github: https://github.com/Joaggi/Prototype-of-a-distributed-language-for-agents

Posters

Forest Height Estimation using SAR Tomography AGU, 2024 Washington DC, US Dec. 2024

Forest Type Classification using SAR Tomography NeurIPS 2024 Vancouver, CA Dec. 2024

Forest Height Estimation using SAR Tomography NeurIPS 2024 Vancouver, CA Dec. 2024

Exploring DINO: Emergent Properties and Limitations for Synthetic Aperture Radar Imagery NeurIPS 2023 New Orleans, USA Dec. 2023

Presentation of research in self-supervised learning
Winner of travel award to present work

Fewshot learning on global multimodal embeddings for earth observation tasks NeurIPS 2023 New Orleans, USA Dec. 2023

Presentation of research in self-supervised learning

Exploring Generalisability of Self-Distillation with No Labels for

SAR-Based Vegetation Prediction NeurIPS 2023 New Orleans, USA Dec. 2023

Presentation of research in self-supervised learning

Large Scale Masked Autoencoding for Reducing Label Requirements on SAR Data NeurIPS 2023 New Orleans, USA Dec. 2023

Presentation of research in self-supervised learning

LEAND: Quantum Latent Density Estimation for Anomaly Detection (Student Abstract) Association for the Advancement of Artificial Intelligence AAAI 2023 Washington DC, USA Feb. 2023

Presentation of research in anomaly detection
Winner of travel award to present work

Doctoral Consortium Association for the Advancement of Artificial Intelligence AAAI 2023 Washington DC, USA Feb. 2023

Presentation of doctoral research

Anomaly Detection through Density Matrices and Kernel Density Estimation (AD-DMKDE) LatinX in AI Research Workshop co-located with the Thirty-Third Neural Information Processing Systems (NeurIPS) New Orleans, USA Nov. 2022

Presentation of research in anomaly detection
Winner of travel award to present work

Robust Estimation in Reproducing Kernel Hilbert LatinX in AI Research Workshop co-located with the Thirty-Third Neural Information Processing Systems (NeurIPS) Vancouver, Canada Dic. 2019

Presentation of research in robust estimation
Winner of travel award to present work

Robust Estimation in Hilbert and Krein Spaces with Reproducing Kernel Fundación COPEC-UC - International Seminar on Artificial Intelligence Sant. de Chile, Chile Nov. 2018

Presentation of master's thesis in research seminar on artificial intelligence

Honors and Medals

Best Paper Award Neurpis 2023 - Climate Change Frontier Development Lab sponsorship by NASA and ESA US 2023

Selected as one of the faculties for developing a height forest estimation using SAR images Frontier Development Lab sponsorship by NASA and ESA Europe-US 2024

Best Young Entrepreneurship Junior Chamber International Colombia 2024 Meritorious Distinction PhD Thesis National University of Colombia Colombia 2023

Selected as one of the researchers for developing a generalizable SAR image method Frontier Development Lab sponsorship by NASA and ESA Europe-US 2023

Selected as one of the 17 best entrepreneurs of Colombia Young Leaders of the Americas Initiative fellowship U.S.A 2020

Generation 20 extension winner Start-Up Chile Chile 2018 Generation 20 winner Start-Up Chile Chile 2018

Third place graduation in computer and systems engineering National University of Colombia Bogotá, Colombia 2014

First place industrial engineering graduation National University of Colombia Bogotá, Colombia 2014

Best Saber Pro in Computer and Systems Engineering Colombian Higher Education Examination Bogotá, Colombia 2013

Best Saber Pro Industrial Engineering Colombian Higher Education Examination Bogotá, Colombia 2013

2nd Place National Programming Olympics Bogotá, Colombia 2006

4th Place National Programming Olympics Bogotá, Colombia 2005

Best student of the whole school Technical Institute of La Salle Bogotá, Colombia 2004-2007

Research Topics and Skills

Research Topics:

Machine Learning, Computer Vision, Space Science, Anomaly Detection, Quantum Machine Learning, Self-supervised Learning, Robust Statistics, Bayesian Statistics

MlOps:

MlFlow, DVC, PyTorch, Tensorflow, Jax, Keras, Sklearn, Pandas, Numpy, Matplotlib, Seaborn, PyPlot, Linux

DevOps:

Google Cloud, AWS, Docker, Kubernetes, Terraform, Jenkins, Git

Back-end:

SpringBoot, Spring Framework, FastAPI, Flask, Django, REST API, R, Matlab

Databases:

PostreSQL, MySQL, MongoDB, Cassandra, HBase, SQlite, SqlAlchemy, Datastore, Hadoop, PySpark

Front-end:

React, Redux, HTML5, LESS, SASS, Javascript, Angular, AngularJs

Programming:

Python, JAVA, Javascript, Kotlin, C, C++, Matlab, LaTeX, Visual Basic

Languages:

Spanish (Native), English (Full-Proficiency), French (B2), Italian(A1)

Reviewer - Program Committee

AAAI- Program Committee US 2025

Reviewer at Machine Learning For Science Workshop 4 Conference Papers US 2024

Reviewer for a grant at The Deutsche Forschungsgemeinschaft (German Research Foundation)

Neural Computing and Applications Self-optimizing processes for multi-frequency PolInSAR image enhancement and analysis Germany 2025 Reviewer at Neural Computing and Applications Multi-Objective Optimization and Predictive Analytics of Strength and Embodied Impacts of CDW-based Geopolymers Using Advanced Machine Learning Techniques US 2025

Jury Machine Learning and Artificial Intelligence Projects Philly Codefest 2025 US - Drexel University 2025

Reviewer at Machine Learning For Science Workshop 4 Conference Papers US

2024 Reviewer at IEEE Geoscience and Remote Sensing Letters 4 Conference Papers US 2024 Reviewer at IEEE Geoscience and Radio Science Conference Paper US 2023

Reviewer at IEEE Geoscience and Radio Science Conference Paper US 2023

Reviewer at Neurocomputings Journal Article US 2023

Jury Digital Entrepreneurship Incubator - Colombo Americano Bogotá, Colombia 2021

Mentor Entrepreneurship Program - Innovation Center Faculty of Engineering University of Santiago de Chile Sant. de Chile, Chile 2020 Mentor and Jury Entrepreneurship Program - GearBox Sant. de Chile 2018

Advisor ‑ Coadvisor

Jinsu Hong Foundation Models and Space Science US 2025-Now
Valmir Moraes Foundation Models and Space Science US 2025-Now
Jasmine Kobayashi Foundation Models and Space Science US 2025-Now
Daniela Martin Foundation Models and Space Science US 2025-Now
Connor, O'brien Foundation Models and Space Science US 2025-Now
Periwal,Mukul Machine Learning base algorithms for shear and bean in gravity of Galaxies US-Drexel 2025 Grace Colverd Tomographic SAR
Reconstruction for Forest Height Estimation US-Europe 2024 Laura Schade
Tree Species Classification using Machine Learning and 3D Tomographic SAR--a case study in Northern Europe US-Europe 2024 [Jumpei
Takami](https://scholar.google.cl/citations?user=Kks1bcIAAAAJ&hl=es) Tree Height Estimation using Machine Learning and 3D Tomographic SAR-a case study in Northern Europe US-Europe 2024
Oscar Alberto Bustos Brinez Development of an algorithm for anomaly detection based on kernel-based density estimation, density matrices and quantum measurements Colombia 2022-2024
Daniela * Martin * Classification of transient astronomical events with recurrent neural networks Colombia 2020-2024
Juan Felipe Osorio Ramirez On * the performance of Kernel Density Estimation using Density Matrices Colombia 2022
Juan Leonardo Padilla Gómez Generative Adversarial Networks through Density Matrices Colombia 2022

Extracurricular Academic Activities

Machine Learning How to win a Kaggle competition Coursera 2020
Machine Learning Improving Deep Neural Networks Coursera 2020 Software
Development Front-End Web Ui Frameworks and Tools: Boostrap 4 Coursera
2020 Machine Learning Bayesian Statistics: From Concepts to Data
Analysis Coursera 2020 Machine Learning Sequence Model Coursera 2020
Machine Learning Convolutional Neural Networks Coursera 2020 Machine
Learning Structuring Machine Learning Projects Coursera 2020 Machine
Learning Neural Networks and Deep Learning Coursera 2020 Software
Development Kotlin for Java Developers Udemy 2019 Machine Learning Curso
de reinforcement learning dictado por Microsoft Edx 2019 Machine
Learning The analitic edge Coursera 2016 Software Security Software
Security Coursera 2016 Software Security Usable Security Coursera 2016
Software Development Software process and agile practices Coursera 2015
Machine Learning The Data Scientist's Toolbox Coursera 2015 Machine
Learning Pattern Discovery in Data Mining Coursera 2015 Machine Learning
Massive Data Mining Coursera 2014 Machine Learning Machine Learning
Coursera 2013 Machine Learning Learning from data Edx 2013 Machine
Learning Introduction to statistics Edx 2013 Machine Learning
Introduction to Statistics Coursera 2013 Computer Science Introduction
to Theoric computation Udacity 2013 Machine Learning Model thinking
Coursera 2012 Universidad Nacional De Colombia Italian language elective
course Bogotá, Colombia 2011 Language English iTEP international test
test of english proficiency Certificate level B2 Bogotá, Colombia 2011
Language English level B2 T&T teaching and tutoring college of colombia
Bogotá, Colombia 2011 Instituto Técnico Central de la Salle III
mathematics meeting Liceo Brother Miguel de la Salle Bogotá, Colombia 2005