Mapping physical spaces for robots
I'm working on the layer between robots and the real facilities they operate in.
My focus is semantic mapping for hospitality back-of-house — turning rooms, equipment, doors, carts, shelves, and operational constraints into structured maps that robots and operators can both reason about.
I'm also exploring how robotic workflows connect back into existing systems of record, starting with hospitality PMS and operations tools.
The goal: make physical AI legible to the systems hotels already use — unlocking self-operating buildings, and eventually abundant shelter.
Current focus
- 01
Semantic site maps
Structured maps of facilities — geometry, equipment, task zones, robot workspaces, and operational constraints.
- 02
Workflow context
Representing what a space means, not just where objects are: where work starts, where handoffs happen, what zones stay clear.
- 03
Systems of record
Connecting robot events back to PMS, task management, inventory, and back-of-house tools.
- 04
Simulation and replay
Using mapped environments for simulation, replay, debugging, and evaluation before workflows reach production.
Why hotels
Back-of-house workflows are structured, operationally dense, and connected to existing systems. Laundry is the first workflow I'm exploring — physical repetition, inventory movement, and real operational constraints. From there the same mapping layer may extend to room operations, kitchen workflows, storage, and maintenance.
If relevant
I'd like to talk with people building or deploying robots in real environments — especially around facility mapping, semantic scene understanding, back-of-house workflows, deployment ops, or PMS integration.