At Agrisoft, I contributed to ML-enabled product features and Python services for AgroScope, an agricultural analytics platform. I built a dataset and model portal that connects image collections, annotation tasks, dataset versions, and model artifacts.
I also connected model inference workflows and Copernicus data services to the Django-based product, developing API and client features for satellite and UAV imagery.
At the Institute of Information and Computational Technologies, I developed and evaluated computer vision approaches for crop and weed detection, plant classification, segmentation, and growth-stage analysis. I created a FastAPI service for batch inference with ONNX models and investigated validation reliability across fields and seasons.
My work also includes an experimental FastAPI service for NDVI GeoTIFF imagery and contributions to ten grant-funded research projects. I co-authored publications on precision agriculture, UAV imaging, environmental monitoring, and geospatial machine learning.
