Building next-generation weather forecasting systems through data-driven machine learning
Ingesting and cleaning forecasting data from weather and observational datasets. Exploring, validating, and preparing high-quality datasets for machine learning.
Building baseline nowcasting models and refining forecast workflows. Experimenting with regional forecasting and severe-weather nowcasting systems.
Developing evaluation and validation pipelines. Producing forecast outputs and performance reports with hands-on operational forecasting experience.
A structured approach to building the future of weather forecasting
Foundation & Data
Set up GPU compute infrastructure, Jupyter Notebook environments, and data engineering workflows on remote servers.
Explore weather and observational datasets, validate source quality, clean and normalize records to produce high-quality validated datasets for machine learning.
Build baseline analytical and predictive models to analyze spatial coverage, temporal patterns, and metadata to organize large datasets.
Modeling & Evaluation
Build and refine nowcasting models using prepared datasets. Experiment with regional forecasting and rainfall or severe-weather nowcasting workflows.
Develop domain-specific forecast evaluation pipelines and understand how datasets shape forecast behavior.
Produce forecast outputs and performance reports. Execute large training jobs on dedicated GPU servers and A100 clusters.
Building next-generation weather forecasting systems through collaborative research and innovation
Build workflows for ingesting and cleaning forecasting data from multiple weather and observational datasets
Create baseline nowcasting models for regional forecasting and severe-weather prediction
Develop evaluation and validation pipelines to assess forecast accuracy and performance
Produce forecast outputs and performance reports with hands-on operational forecasting experience