PhD candidate in Atmospheric and Oceanic Science at the University of Maryland working at the intersection of Earth system predictability, machine learning, and causal inference.
I develop and apply data-driven methods to understand and predict variability in the climate system, with interests spanning:
- 🌎 Subseasonal-to-seasonal predictability
- 🌪 Large-scale circulation and extreme events
- 🤖 Machine learning for Earth system science
- 🔍 Explainable AI and causal discovery
- 📈 Forecasting under uncertainty
I’m especially interested in combining physical understanding with statistical and machine-learning methods to tackle challenging prediction problems in complex geophysical systems.
Current/recent projects include:
- TraCE-ST — tracing the space-time causal origins of Earth system extremes
- Weather regime predictability using interpretable ML
- Causal discovery for Earth system dynamics
- Applied forecasting for air quality, energy, and environmental risk
Mostly working in:
Python | xarray | numpy | pandas | scikit-learn
PyTorch | XGBoost | Tigramite | Cartopy
HPC workflows | Earth system model analysisRepositories related to:
- Research code and reproducible workflows
- Machine learning for weather and climate
- Causal inference experiments
- Visualization and Earth system diagnostics
- Side methodological projects
I care about reproducible, interpretable, and scientifically grounded machine learning.
- 🌐 Website: https://jhayron-perez.github.io/
- Scholar: https://scholar.google.com/citations?user=lzLdYv0AAAAJ&hl=en&authuser=1
- 💼 LinkedIn: https://www.linkedin.com/in/jhayron-s-p%C3%A9rez-carrasquilla-b7930321a/
- 📫 Email: jhayron@uchicago.edu; jspecar@gmail.com
Always happy to connect around climate dynamics, prediction, ML, and causal methods.