Ship fast. Tight feedback loops. Clear systems.
Independent senior engineer for maps, geospatial, and data systems. I build the v1, untangle what's tangled, and make Postgres and PostGIS fast — Ruby on Rails, Go, Python, H3. Based in Alaska, remote worldwide.
What I can help with
- Build or fix your map stack — spatial APIs, vector tiles, geocoding and place search on Postgres / PostGIS and H3, fed by Overture Maps, OpenStreetMap, or your own GeoJSON.
- Decide buy vs build — a straight call on when Mapbox, ArcGIS, CARTO, Felt, Geoapify, or Stadia Maps is enough and when custom pays off, then the implementation either way.
- Build your v1 — an idea or designs, taken to a real product in production. Ruby on Rails, Go, or Python on Postgres.
- Untangle a messy system — tangled logic, slow reports, data nobody trusts, replaced with something clean.
- Tune performance — deep SQL / Postgres / PostGIS: large-scale data, slow queries made fast.
- Make data pipelines hold up — ingest, dedup, and serve large datasets with minimal infrastructure: Postgres, DuckDB, Python on AWS, DigitalOcean, or Cloudflare.
Selected work
- Dataplor — built and scaled a global GIS for place dedup and brand–place analysis at 50B+ geometries and 10B+ place observations.
- Mothership — untangled dispatch logic onto one Postgres system and cut $10k+/day in routing costs.
- iOverlander — backend, infra, and tile pipeline rebuilt and still shipping; engagement has run async-only (no calls or meetings) for two years.