Projects
Blazars Population ML | Jan 2026 – Present
- Applying classical ML and Bayesian inference to the Fermi-LAT 4LAC-DR3 catalog (3,407 sources) to recover physical jet parameters from multi-wavelength observational data.
- Building a neural network to learn the mapping from physical jet parameters to observable SEDs accurately enough to compare with Tleco results.
- Developing a forecasting model for blazar gamma-ray flares using a general-purpose time series foundation model, fine-tuned with Tleco-generated synthetic light curves.
Used: Git, Python: Numpy, Pandas, Matplotlib, Scipy, Astropy, Scikit-learn, PyTorch, Pytest, Jupyter
Wind and Spores | Aug 2024 – Present
- Open source R & C++ package for dispersion of spores in hilly terrains.
- Implemented Lagrangian particle dispersion, atmospheric turbulence, and fluid flow over complex hilly terrain with forest canopy, based on analytical models.
- Parallelized particle tracking using RcppParallel and OpenMP. Designed for scalability to tens of thousands of particles across kilometer-scale domains.
Used: C++, OpenMP, Git, R: RStudio, Rcpp, ggplot2, dplyr, tidyr, RcppParallel
Tleco | Jan 2024 – Sep 2024
Tlecostands for both in the fire and rise in the Nahuatl language.- Open Source Code for Simulation of Relativistic Particles Radiation
- Numerical code that simulates particles in relativistic plasma, and the rise of radiation from accelerating particles.
- Consists of both Rust functions and Python functions previously built in the Fortran code
Paramo.
Used: Rust, HDF5, Git, Python: Numpy, Matplotlib, Scipy
Publications using Tleco
- Davis, Z., Rueda-Becerril, J.M., Giannios, D. (2024). “Tleco: A Toolkit for Modeling Radiative Signatures from Relativistic Outflows”. The Astrophysical Journal, 976(2), 182, 12. DOI:10.3847/1538-4357/ad8bc2
Paramo | Oct 2018 – Apr 2022
Open Source Code for Radiative Transfer Simulations in Relativistic Astrophysics
- Independently developed production-grade HPC code for Monte Carlo radiative transfer simulations.
- Optimized through systematic profiling, algorithmic improvements, and memory efficiency enhancements.
- Applied gradient descent optimization to fit simulation parameters to NASA telescope observations.
- Used by 3+ graduate students; resulted in 5+ peer-reviewed publications.
- Developed comprehensive Python analysis and visualization pipeline.
Used: Fortran, HDF5, OpenMP, MPI, Git, Python: NumPy, Pandas, Matplotlib, SciPy
Publications using Paramo
- Rueda-Becerril, J.M., Harrison, A.O., Giannios, D. (2021). “Blazar jets launched with similar energy per baryon, independently of their power”. Monthly Notices of the Royal Astronomical Society, 501(3), 4092-4102. DOI:10.1093/mnras/staa3925
- Rueda-Becerril, J.M., Harrison, A.O., Giannios, D. (2021). “The blazar sequence revised”. Astronomische Nachrichten, Volume 342, Issue 1-2, pp. 147-152. DOI:10.1002/asna.202113895
- Rueda-Becerril, J.M. (2021). “A numerical approach for radiative cooling in relativistic outflows”. Astronomische Nachrichten, Volume 342, Issue 1-2, pp. 277-282. DOI:10.1002/asna.202113919
- Davis, Z., Rueda-Becerril, J.M., Giannios, D. (2022). “Balancing turbulent heating with radiative cooling in blazars”. Monthly Notices of the Royal Astronomical Society, 513(4), 5766-5779. DOI:10.1093/mnras/stac1282
- Combi, L., Siegel, D.M. (2023). “GRMHD Simulations of Mergers with Weak Interactions: r-process Nucleosynthesis and Electromagnetic Signatures of Dynamical Ejecta”. The Astrophysical Journal, 944(1), 29. DOI:10.3847/1538-4357/acac29
Portfolio — practice projects, course work, and reusable templates
Machine Learning From Scratch
Fundamental ML algorithms implemented from scratch to demonstrate deep understanding of underlying mathematics:
| Algorithm | Description |
|---|---|
| Neural Network | Feedforward NN solving XOR problem with backpropagation |
| Decision Tree | Classification tree with Gini impurity |
| Random Forest | Ensemble with bootstrap aggregating |
| Gradient Boosting | Sequential ensemble with residual fitting |
| Movie Recommender (WIP) | Collaborative filtering (user-based, item-based, matrix factorization) |
Used: Python: NumPy, Matplotlib, Scikit-learn
DataCamp
Applied data science projects completed through DataCamp courses:
| Project | Topic | Description |
|---|---|---|
| Avocado Toast Analysis | Exploratory Data Analysis | Price and trend analysis of avocado-related datasets with visualization |
| Nobel Prizes | Historical Data Analysis | Examining Nobel Prize award data, winners, and trends over time |
| Oldest Businesses | Business & History | World’s oldest businesses: longevity, characteristics, and geographic distribution |
| Sleep Data Analysis | Health & Predictive Modeling | Sleep disorder prediction and health metrics analysis |
Used: Python: Pandas, NumPy, Matplotlib, Seaborn
Kaggle
| Project | Competition / Dataset | Description |
|---|---|---|
| Titanic — A Study of a Shipwreck | Titanic: Machine Learning from Disaster | Binary classification to predict passenger survival; female survival rate 74.2% vs. male 18.9% |
| Telco Customer Churn | Telco Customer Churn | Predict which telecom customers are likely to churn; EDA + model comparison across 7 classifiers on 7K records |
Used: Python: Pandas, Scikit-learn (Logistic Regression, KNN, Decision Tree, Random Forest, Gradient Boosting, SVM), XGBoost
Templates
Reusable Jupyter notebook templates for standard data science workflows, available in the Templates directory:
| Template | Purpose |
|---|---|
| Data Cleaning | Data quality checks and cleaning workflows |
| Data Querying | Data retrieval and querying operations |
| EDA | Exploratory data analysis and visualization |
| ML Preprocessing | Feature engineering and preprocessing for ML models |
Tonalpowalli (WIP)
I have a GitHub repository for my tonalpowalli (the mexihkah count of the days) project (in progress).
Gists
I have a collection of GitHubGists that you can explore.