NASA Goddard Space Flight Center
|Assistant Research Scientist
Greenbelt, Maryland, US
Summary
Leads advanced research in AI-driven Earth observation and climate modeling, developing and deploying high-performance computing solutions for critical environmental assessments.
Highlights
Architected and deployed AI/ML frameworks across NASA's 32-GPU distributed computing infrastructure, accelerating global methane emission reconciliation by 40% and delivering operational 1km-resolution greenhouse gas flux estimation by integrating 7 diverse satellite datasets.
Developed GeoCryoAI physics-informed transfer learning framework, reducing permafrost carbon feedback prediction uncertainty by 39% (RMSE from 1.997cm to 1.007cm) and achieving 96.4% detection accuracy across billions of observations.
Quantified permafrost zero-curtain phenomena, revealing its control over up to 40% of the remaining carbon budget for the 1.5°C warming limit by processing 54+ million observations on a 32-GPU distributed infrastructure.
Designed a scalable data engineering framework for NASA's NISAR satellite mission, establishing operational circumarctic monitoring with 30m resolution and enabling 3-6 month forecasts from multi-billion-row geospatial databases.