Remote sensing · PostGIS · Decision support
Earth Observation
Geospatial engineering for agriculture and the environment — satellite data, PostGIS pipelines and decision support that reaches the people doing the work.
Maps are how I think. My route into software was through a geospatial surveying firm, where the job was to make railway engineering data legible to the people building the railway. From there I moved into agriculture, where the same question keeps coming back in a different costume: how do you turn a very large amount of spatial data into one useful decision?
Most of my work sits behind the map rather than on it: data models, ingestion pipelines, validation, tiling, and APIs that make the map fast and honest.
What I bring
Satellite data without the mystique
Sentinel-2 and Landsat processing, cloud masking, vegetation indices and time-series analysis. I care as much about the provenance of a pixel as its value.
PostGIS and pipelines
Vector and raster data models, spatial indexing, tiling, and scheduled ETL that can be re-run safely. GDAL, rasterio, GeoPandas and a healthy respect for coordinate systems.
Decision support, not dashboards
Tools built around an actual decision: when to irrigate, where to sample, which field needs attention first. A chart is only useful if it changes what someone does.
Data products that stay up
Clear SLAs, caching, monitoring and graceful degradation when a provider is late or a scene is cloudy. Earth observation is weather-dependent; the software should not be.
Selected work
Yield prediction for growers
Backend services that ran crop models and disseminated their output to agronomists and growers. The interesting work was not the model itself but making its uncertainty understandable and its recommendations actionable.
Field boundaries and crop classification
Pipelines for deriving and validating field geometries from imagery and survey data, with a review workflow that let humans correct what the model got wrong.
Flood and land-cover monitoring
Prototype systems for tracking surface water and land-cover change across seasons, giving analysts a repeatable view instead of a folder full of one-off exports.
Engineering maps for Irish Rail
Client-facing web maps for railway surveying projects: asset layers, project extents, and a shared spatial language for crews, engineers and clients.
Toolbox
How I think about maps
A map is an argument. It selects, simplifies and ranks the world. The job is to make that argument explicit, defensible and useful.
That usually means showing the uncertainty, keeping the raw data reachable, and designing for the person standing in a field with one bar of signal. It also means remembering that the most important layer is often the one someone drew by hand.
Let's talk
If you are working on agriculture, climate, infrastructure or anything else with a spatial question at its centre, send me an email. I am always happy to talk about maps, pipelines and the strange places data comes from.