GIS & Geospatial Engineering

I build maps that survive production.

Every map demo looks great; the trouble starts when real users and real data show up. Geospatial has been a recurring thread through my 20+ years of shipping software — the Esri/ArcGIS stack on FM Global’s property-risk analytics platform, a greenfield multi-tenant subsurface-mapping SaaS I architected on Azure, and map products going back to a Google Maps healthcare app I wrote when that API was still young. Geocoding, storage schemas, platform integration, the viewer on top — I’ve personally owned every one of those layers, and this page is how I think about the work.

How I Work with Geospatial

Six disciplines I’ve practiced hands-on, with production map data on the line every time.

Esri / ArcGIS Platform Engineering

I learned early that Esri behaves best when you refuse to treat it as a desktop tool. ArcGIS Enterprise and ArcGIS Pro, arcpy Python automation, Workflow Manager, Experience Builder, geoprocessing/GP services, and Arcade — my rule is that ArcGIS gets no special exemptions. It goes under version control, automation, and documentation like every other production system I run.

Geocoding & Location Data Quality

No geocoder is right everywhere, which is why I stopped trusting any single one. I once put five geocoders head-to-head against the same address set to work out which blend a platform should actually rely on. Address deduplication, parcel data, and the unglamorous quality plumbing that determines whether a pin lands on the right building or the one next door — that plumbing is where I spend my time.

Geospatial Data Architecture

The schema underneath the map decides what the product can ever become, so that’s where I start every design. GeoJSON and Parquet layouts, KML/KMZ ingestion pipelines, and PostgreSQL/PostGIS-style modeling for lines, polygons, markers, layers, features, and tracks — built so the storage still holds up after the feature counts get serious.

Map-Driven Web & Mobile Products

A field tool has to hold up through a full workday, not a demo. Before committing a platform to a base map, I settled the Mapbox vs Google Maps question with actual pricing math instead of preference — and I designed offline mode after accepting a hard truth: field crews lose signal exactly where the work happens.

Geospatial Processing at Scale

I’ve pushed satellite imagery-derived data through Databricks and Spark, kicked the tires on Apache Sedona, Shapely, and Rasterio myself rather than reading the docs and guessing, and compared commercial sources like Ecopia and Precisely. The lesson that stuck: once the geometry outgrows one machine, you’re building a pipeline, not a map.

GIS Security & Governance

A latitude and longitude can be somebody’s home or a client’s critical asset, so I treat location data as sensitive by default. I wrote the persona/permission authorization matrix for an Esri Workflow Manager layer myself and put GIS APIs behind WAF and API management — the map platform has to clear the same security review I hold everything else to.

5
Geocoders I benchmarked head-to-head before recommending a blend for a property-risk analytics platform
50
Developers I led on a satellite-imagery geospatial analytics platform
6
Map-object types I modeled for a multi-tenant geospatial SaaS — lines, polygons, markers, layers, features, tracks
20+
Years I’ve been shipping production software underneath these maps

Where I’ve Shipped Maps

Three platforms — insurance, utilities, healthcare. Each one made it to production, and each one taught me something.

Insurance · Esri & Databricks

FM Global — Property-Risk Geospatial Platform

As Solutions Architect on the property risk analytics platform at one of the world’s largest commercial property insurers, I owned the geospatial data architecture end to end. That meant running the five-geocoder evaluation myself, designing the Parquet/GeoJSON storage schemas, and going deep across ArcGIS Enterprise, ArcGIS Pro, arcpy, and Workflow Manager. Getting ArcGIS Pro under real CI/CD wasn’t in anyone’s playbook, so I wrote that process documentation from scratch — along with the platform’s Esri authorization matrix.

Utilities · Azure SaaS

Multi-Tenant Subsurface-Mapping SaaS

A national subsurface-mapping and utility-locating company brought me in as Solutions Architect on a greenfield multi-tenant geospatial SaaS on Azure. I designed the map-object microservice APIs (lines, polygons, markers, layers, features, tracks), worked out the KML/KMZ/GeoJSON-to-database ingestion and schema, and built the multi-tenant sharing and permissions model. The hardest calls were offline mode for field crews — connectivity dies right where locate work happens — and the Mapbox vs Google Maps decision, which I settled with real pricing analysis while also leading an offshore delivery team.

Healthcare · Google Maps

Provider-Visit Tracking Application

Well before mapping was fashionable, I wrote a Google Maps-based web application that tracked in-home healthcare provider visits for anti-fraud verification — plotting each visit so the paper record could be checked against where the provider actually went. That project taught me that a map is an audit tool as much as a picture, and I’ve been pulling on that thread ever since.

Does your data keep asking where?

If there’s a where hiding in your data, I can build the system that answers it — Esri platform work, geocoding pipelines, the storage schema underneath, and the map product on top. I work remote, Corp-to-Corp.