About
Sarah Ilyas
I am an ML Engineer based in Istanbul, building production forecasting systems and writing about MLOps, time series ML, and uncertainty quantification.
Currently working on Cropcast — an agricultural yield forecasting pipeline that detects model drift automatically and retrains itself every week without human intervention. The system tracks 6 crops across 13 countries and caught a 1334% MAE degradation on the Grapes model automatically.
Previously worked on data intelligence systems for agricultural export tracking across Mexico and Latin America.
Skills
PythonXGBoostProphetSHAPOptunaMLflowFastAPIStreamlitDockerAWS EC2GitHub ActionsPrefect