Richard I.

Geospatial Data Engineer | Production ETL & Remote-Sensing Pipelines

I build production-grade data pipelines for location data, the kind that survive real scale, messy sources, and senior technical review. Most data engineers can't work in GIS, and most GIS specialists can't ship reliable pipelines. I do both, which is exactly what location-heavy projects need. If you're working with geographic data, satellite or aerial imagery, parcel and infrastructure records, point-of-interest datasets, anything tied to coordinates, I turn it into clean, queryable, automated systems instead of one-off scripts that break the moment the source changes. What I do: - Geospatial ETL pipelines (PostGIS, geopandas, rasterio) that ingest, clean, and structure spatial data at scale - Remote-sensing and imagery workflows, sourcing, tiling, and organizing aerial/satellite data across many locations - Web scraping and data extraction (Playwright, BeautifulSoup) with proper deduplication, rate-limit handling, and validation - Config-driven, schema-adaptive pipelines that adapt to new sources without a rewrite - Spatial APIs and mapping backends (FastAPI, PostgreSQL/PostGIS, Leaflet) - AI-connected spatial systems (MCP servers that expose PostGIS queries and geospatial tools to AI assistants) How I work: I diagnose the real bottleneck before writing code, I'm honest about tradeoffs (including when the hard part is data cost or source limits, not engineering), and I build for handoff, documented, maintainable, and yours. Proof: - Processed 6M+ US building permit records through a config-driven ETL pipeline (ConstructIQ) - Built a nationwide vendor directory of 4,400+ records via multi-phase scraping and enrichment (EventStarted) - Mapped fiber infrastructure across Lagos State in PostGIS (UDIGAP) - Built hybrid MCP servers exposing PostGIS and standalone geospatial tools to AI assistants, published to Glama (geo-mcp) - B.Sc. in Surveying & Geoinformatics — the geospatial fundamentals behind the engineering Tell me what your data looks like and what you need out of it, and I'll map the approach against your constraints before we talk price.