Smart Spatial System (s3geo)
Ask a spatial question in English or Persian — get a validated, traceable workflow, map layers and reports.
Open to meaningful engineering work
Urmia, Iran · 37.5527°N 45.0761°E
Geospatial AI Engineer · Backend Engineer
I build natural-language spatial analysis systems, high-performance geospatial APIs and managed PostGIS infrastructure — turning maps, data and LLMs into reliable software.

How s3geo turns a question into a validated spatial workflow.
Ask a spatial question in English or Persian — get a validated, traceable workflow, map layers and reports.
Managed cloud PostGIS — a ready database in under a minute.
A plugin-based GeoAI kernel and Python SDK for parsing, routing, planning and executing natural-language spatial queries.
A Persian map assistant that answers plain-language questions with places, routes and spatial results.
Reproducible studies that stress-test s3geo on real cities and real data.
Can an LLM, with no hints, plan the same spatial analysis an expert would? A reproducibility study of health-facility accessibility.
20/20 unhinted runs · Jaccard 1.0
A raster time-series pipeline for vegetation change, optimised without changing a single result.
~20× faster · 18.5 min → 55 s
The reference study: an accessibility ranking executed end to end.
Tracking the shrinkage of Lake Urmia from remote-sensing time series.
Spatial analysis of the local real-estate market.
Transit-oriented development gradient across Tehran — paper in draft.
Natural-language interfaces for spatial data: LLM planning with pre-execution validation, traceable outputs and reproducible results.
Model Context Protocol servers (Tools, Resources, Prompts) that let AI agents work safely with your GIS stack.
Scalable Python/FastAPI/Django services over PostGIS, OGC standards (WMS, WFS, WCS, WMTS) and OpenAPI-documented REST.
Schema design, indexing, performance, backups — or a managed instance on postgis.ir.
Tell me about the problem, the data and the timeline. I read every message.