Welcome to the new Python version of the site. This new version is the result of an experiment in which I asked Claude Code to completely rewrite my original website and backend code, converting from MATLAB to Python. As you can see, it successfully rewrote everything, including the backend pipelines, as well as added GEFS forecasts to several of the plots. This site will eventually replace the original once I sunset the old servers and continue to experiment with agentic AI.

This is still under development so there might be some plotting errors.

As always, send feedback to kyle at macritchie.me

Kyle MacRitchie

Data Scientist, Meteorologist, Geospatial Scientist
kyle at macritchie dot me
github
LinkedIn
Washington, DC Area
Kyle MacRitchie

I excel in harnessing geospatial satellite, weather, and climate data to create sophisticated machine learning models, delivering precise forecasts for agricultural and energy sectors.

Complete Resume in PDF form: click here.

Work Experience

Senior Meteorologist

DV Trading | 2024 - present
  • Focus on agricultural forecasting.

Applied Scientist

Descartes Labs | 2021 - 2024
  • Employed a tech stack (scikit-learn, PyTorch, boto3) on AWS to develop, train, backtest, and deploy ML models and pipelines for forecasting yield and production of key crops (e.g., soybeans, sugar, robusta & arabica coffee) throughout the growing season.
  • Designed advanced weather prediction models with PCA regression, vision transformers, and attention techniques in PyTorch, producing accurate subseasonal to seasonal forecasts crucial for crop modeling.
  • Created a tropical cyclone prediction system that delivers calibrated probabilistic forecasts for rain and wind impacts from tropical cyclones over specific regions, up to four weeks in advance.
  • Worked directly with clients to understand their needs, incorporate feedback, and ensure maximum benefit from our products.
  • Led pivotal projects, including the development of an internal weather API, to improve access to weather data across the company.

Meteorologist

Climate Prediction Center (NWS) | 2018 - 2021
  • I developed forecast models to predict climate features using recurrent neural networks in a Tensorflow framework.
  • I created and applied statistical techniques to bias correct and calibrate ensemble model forecasts.
  • I created several ArcPy scripts that allowed us to use ArcMap in our operational products.
  • I created, implemented, and maintained several daily pipelines that were critical to 24/7 forecast operations.
  • I mentored and led contract scientists.

Peer Reviewed Publications and Presentations

  • MacRitchie, K., and C. Schreck. 2021: Tropical intraseasonal variability [in "State of the Climate in 2020"]. Bull. Amer. Meteor. Soc., 102 (8), S210–S213
  • L'Heureux M., E. Becker, M. S. Halpert, Z.-Z. Hu, K. MacRitchie, and M. Tippett. 2021: ENSO and the tropical Pacific [in "State of the Climate in 2020"]. Bull. Amer. Meteor. Soc., 102 (8), S205-S210
  • Liu, Z, D. Ostrenga, B. Vollmer, B. Deshong, K. MacRitchie, M. Greene, and S. Kempler. 2017: Global precipitation measurement mission products and services at the NASA GES DISC. Bull. Amer. Meteor. Soc., 98, 437-444.
  • MacRitchie, K., and P.E. Roundy. 2016: The two-way relationship between the Madden Julian oscillation and anticyclonic wave breaking. Quart. J. Royal Meteor. Soc., 142, 2159-2167.
  • Roundy, P.E., N. Sakaeda, K. MacRitchie, L. Gloeckler, 2017: Weather-climate interactions and MJO influences. Climate Extremes: Patterns and Mechanisms, S.-Y.S. Wang et al., Eds., Amer. Geophys. Union, 139-163.
  • MacRitchie, K., and P. E. Roundy, 2012: Potential vorticity accumulation following atmospheric Kelvin waves in the active convective region of the MJO. J. Atmos. Sci., 69, 908-914.
  • Roundy, P. E., K. MacRitchie, J. Asuma, T. Melino, 2010: Modulation of the global atmospheric circulation by combined activity in the Madden–Julian oscillation and the El Niño–Southern oscillation during boreal winter. J. Climate, 23, 4045–4059.
  • Invited speaker: 18th Annual Maxar Energy Conference, October, 2019 on the MJO and ENSO at CPC.