51.0543° N · 3.7174° E — Open-source Python

Loucas
Diamant-Boustead

Geospatial systems & AI for the environment

I build optimization models, machine-learning pipelines and web tools — not just maps. Turning spatial data into environmental decisions, entirely in open-source Python, from architecture scale to the scale of a coastline.

What I build

Spatial optimization

Linear & mixed-integer models that allocate land, energy and habitat under hard constraints.

geoAI & remote sensing

Machine learning on satellite & aerial imagery — segmentation, detection and change over time.

Mobility & routing

High-performance transit engines on real timetable data to measure access and modal shift.

Web tools & pipelines

FastAPI services, data pipelines and interactive dashboards that put analysis in people's hands.

A physical geographer who learned to build the tools the questions needed.

My route into geospatial work ran through the field — three Arctic and Subarctic seasons mapping ground and water, and a master's in stable-isotope hydrology spent chasing meltwater across the Mackenzie Delta. That work made one thing obvious: the hard part is rarely the map. It is the system behind it — the model, the pipeline, the optimization that turns a pile of measurements into a decision someone can defend.

So I build those systems. Today I work as one of two geospatial specialists at an architecture and urbanism practice, in a Python-first, fully open-source stack that bridges building scale to city and regional scale. The thread running through all of it is environmental: renewable-energy siting that respects ecology, mobility analysis that nudges cities off cars, remote sensing that tracks the things worth protecting.

I'm based in Ghent, work in English and French, and I'm steadily wearing down Dutch. When I'm not at a terminal you'll find me on the water — a long habit from competitive rowing — or building small side projects that have no business being as over-engineered as they are.

280 GW
of offshore energy modelled across the North Sea
7
countries reconciled in a single planning optimization
50k
origin–destination pairs in transit access analysis
100%
open-source — GeoPandas, PuLP, rustworkx & friends

Selected work

All projects