Zootechnical Engineer and Data Scientist with over 35 years of professional experience in animal production systems, rural development, agricultural policy and public administration. Currently specialising in the application of data science, geospatial analysis, remote sensing and machine learning to environmental monitoring and agricultural policy.
At the Northern Regional Coordination and Development Commission (CCDR-N), Portugal, coordinating the groundwater nitrate monitoring programme for the Esposende–Vila do Conde Vulnerable Zone, under the EU Nitrates Directive.
This work combines data science, geospatial analysis and environmental regulatory compliance:
- Data integration —This integration is part of a broader multi-source effort — including livestock and survey data — that has nearly doubled the universe.
- Deep learning and remote sensing — Development of a U-Net semantic segmentation model to automatically detect greenhouses in satellite imagery, including structures missing from official registers
- Geostatistics — Spatial analysis of nitrate concentrations in groundwater using inverse distance weighting (IDW) and kriging
- Data pipelines and dashboards — Development of Excel-to-PostgreSQL data ingestion pipelines, ArcGIS integration and interactive Streamlit applications for environmental monitoring teams
Independent expert evaluator for the European Research Executive Agency (REA) since 2015, assessing research and innovation proposals under Horizon 2020 and Horizon Europe, with a focus on climate and environmental sustainability, the efficient and circular use of natural resources, the digital transformation of small- and medium-sized farms, and the development of more diversified and resilient livestock production systems; since 2024, also contributing to the review of the implementation and progress of projects funded under Horizon Europe.
Pre-Bologna university degree in Zootechnical Engineering (Animal Science and Engineering), comprising four and a half years of coursework followed by a final internship involving an applied study.
Master’s degree in Tools and Techniques Supporting Rural Development (Mestrado em Instrumentos e Técnicas de Apoio ao Desenvolvimento Rural).
Earlier professional experience focused on small-ruminant production systems and the sustainability of Mediterranean farming within the FAO-CIHEAM research network:
- Author or co-author of more than 50 scientific and technical publications
- Co-editor of CIHEAM's Options Méditerranéennes series
- Member of the scientific committees of international FAO-CIHEAM seminars
- Recipient of the national “Best Regional Information Technician” Award from Statistics Portugal (INE) in 2010, for coordinating the Agricultural Census in the Entre Douro e Minho region
Member of the Portuguese Order of Engineers (Ordem dos Engenheiros), affiliated with its College of Agronomic Engineering, which provides the professional framework for Zootechnical Engineering.
Holder of the professional title of Senior Engineer, internship supervisor appointed by the College of Agronomic Engineering, and member of the Order’s Expert Panel.
Member of the Portuguese Association for Data Science and Artificial Intelligence.
DataCamp Certified Professional Data Scientist.
Python (pandas · GeoPandas · scikit-learn · TensorFlow) · PostgreSQL ·
ArcGIS Pro · Streamlit · Plotly · Git
- Farm-holdings baseline integration — merging the national land-parcel registry with the zone's monitoring platform into a deduplicated, spatially anchored universe of 3,400+ holdings — the denominator for nitrate pressure analysis and the sampling frame for annual farmer surveys
- Nitrate monitoring — Esposende–Vila do Conde Vulnerable Zone — data pipeline and interactive dashboard for a Nitrates Directive monitoring programme: from field Excel records to PostgreSQL, ArcGIS and Streamlit Cloud, with automated monthly publication
- Greenhouse detection with U-Net — deep learning on Sentinel-2 imagery to map protected horticulture across a Nitrates Directive Vulnerable Zone, revealing 1,200+ structures absent from official land registers
- Greenhouse registry verification with ML — classical machine learning on RGB-derived features to audit a decade-old GIS registry against current imagery, with per-polygon confidence to target field inspection
- Livestock stocking rates — joining parcel-registry forage surface with livestock records to compute per-holding stocking rates and their year-on-year evolution — the animal-pressure side of nitrate monitoring