PointFold
Structures the geography.
For teams with location data that need meaningful clusters, groups, and planning zones without a full routing or scheduling model.
Open PointFold ↗GeoHub · Geospatial intelligence products
GeoHub provides practical tools for spatial clustering, service-zone design, route optimization, and workforce scheduling. PointFold structures the geography. RouteForge optimizes the work.
GeoHub is the business-facing geospatial product studio. Aurora is the applied-AI lab for local agents, multimodal systems, and experimental conversational tools.
A focused product suite for teams making decisions from location data: cluster demand, define service geography, and optimize routes, schedules, workers, visits, and constraints.
Local AI agents, multimodal clients, clinical NLP, and experimental conversational systems — developed as applied research and product exploration.
The products address different levels of operational complexity and can be used independently or in sequence.
Structures the geography.
For teams with location data that need meaningful clusters, groups, and planning zones without a full routing or scheduling model.
Open PointFold ↗Optimizes the work.
For teams with workers, visits, time windows, skills, capacity, and constraints that need efficient routes and schedules.
Explore RouteForge ↓The commercial focus is PointFold and RouteForge. The remaining work demonstrates applied geospatial delivery and ongoing AI research.
Two focused tools for the path from location data to operational action.
A spatial workbench for location data.
Upload coordinates, discover natural clusters, define service zones, and export planning-ready spatial structure — all processed locally in the browser with Turf.js.
Open PointFold ↗
An operations workbench for routes and schedules.
Optimize routing, scheduling, and dispatch decisions across workers, visits, skills, capacity, time windows, constraints, and operational scenarios — informed by live and historical traffic plus weather forecasts for more accurate ETAs and sharper operational decisions.
I build local, on-device apps tailored to a specific industry's privacy and field realities — and connect them back to a central GeoHub dashboard for full last-mile coverage, without exposing sensitive data.
An optimized day of visits — computed entirely on the device.
A Flutter app that gives a Personal Support Worker an optimized multi-stop route across their day's home-care visits, with routing, optimization, and map rendering running fully on-device. The visit list and patient addresses never leave the phone — no server, no network call, demonstrable in Airplane Mode. It still syncs to a central GeoHub dashboard for planning and oversight, without exposing patient data — the privacy-first mobile vertical of the NORC last-mile work.
A public-facing example of geospatial analysis translated into a practical decision interface.
Exploring how immersive geospatial environments can support analysis and decision-making.
A Google VR / 3D Tiles geospatial experiment with an LLM assistant for navigating, querying, and interpreting spatial environments.
Secondary-studio work in local AI agents, multimodal clients, clinical NLP, and experimental conversational systems.
An experimental conversational agent interface built around local and self-hosted AI workflows.
Open Aurora Agent ↗
A private iOS conversational client for local and cloud language models.
A multilingual conversational system for real-time speech and text translation.
A tunable multimodal client for text, image, video, and audio workflows across local and hosted providers.
A clinical NLP workspace for evidence-linked risk analysis and image-aware conversational support.
These engagement tiers are placeholders while packaging is finalized. No fixed prices are published yet.
Pricing coming soon
Pricing coming soon
Scoped after discovery
Scoped after discovery
Scoped after discovery
I build practical decision tools where geospatial analysis meets operational planning. GeoHub is the primary commercial studio, with PointFold for spatial structure and RouteForge for operational optimization. Aurora is the secondary applied-AI lab.
A recurring theme in my work is privacy-first, on-device apps built for a specific industry's needs — like the Last-Mile PSW routing app that runs entirely offline — that still connect to a central hub or dashboard for full last-mile coverage without sensitive data ever leaving the device.
A lot of this work runs at the edge: I also build robotics and on-device AI on constrained hardware — ROS 2 perception, navigation, and local LLM/voice stacks on the NVIDIA Jetson Orin Nano (see my JetBot / ROS 2 work). If you're working on robotics or edge AI and want to collaborate, get in touch.
Let's talk
Dashboards, maps, clustering, route optimization, forecasting, or AI assistants — describe the operational problem, and I’ll help scope the right solution.