Mohamed A M Elansary, PhD
Target: Machine Learning Research Engineer — WindBorne Systems
Sourced insights
- AI data assimilation: WindBorne AI DA closes the loop on Global Sounding Balloon observations → better initial conditions; publicly reported 3.5% more accurate than ECMWF physics-based DA at 12h for Z500 — assimilation is the product thesis, not forecast demos alone. Source: windbornesystems.com/blog/windborne-ai-da
- JD ML RE seat: AI-based data assimilation of balloons, satellites, and weather stations into forecasts; progress toward a weather foundation model; clean/combine messy multi-source weather datasets; rapid experiments → reusable research infrastructure. Source: windbornesystems.com/careers/machine-learning-research-engineer
Proof — assimilation × UQ/eval × geospatial/time-series × ship
- PhD Environmental Engineering, TAMUK 2022: multi-basin, multi-hydroclimate surface-water/groundwater forecast uncertainty quantification & reduction on HPC (MODFLOW, VIC, PIHM, NASA LIS) over USGS/NOAA/NASA stacks — assimilation-adjacent multi-source obs → forecast skill.
- Geospatial + time-series: ArcGIS/QGIS 8 years; AMS 2021 spatio-temporal floods & droughts; multimodel streamflow forecasts across hydroclimates (AMS 2018).
- Messy multi-source Earth/hydro data fluency + publication-grade QA — the same clean/combine/decide-usefulness muscle the WindBorne seat needs for balloons/satellites/stations.
- Ships production Python AI systems (Vertexium agentic LLM + retrieval + multi-tenant agents; EDAT GRI/SASB/TCFD) — experiments that become durable systems.
- Honest frame: not claimed WeatherMesh/AI-DA owner. The PhD who ships. Bay Area OK · Prefer take-home · $140k–$240k + equity story.