WFM Twin
WFM Twin — Robot workforce intelligence

Put capable robots to work.

WFM Twin is building the workforce-management and evidence layer for robotics—matching available machines to work they are capable, authorized, and economically justified to perform.

Today

A simulated loop

Isaac Sim + ROS execution with linked task and cleanup records.

Building next

The decision layer

Reusable task, robot, certification, authorization, and evidence contracts.

Looking for

Sharp collaborators

Robot owners, integrators, operators, and safety or domain experts.

Eligibility trace / FOL-006

Simulated decision record

Replay ready

Observation

Dense surface fuel

Task

Selective clearance

Owner policy

Authorized

Robot

Fuel-Ops-01

Decision

Eligible → assign supervised run

Evidence plan attached

Preview data is illustrative. Fuel Ops Loop currently runs as a scripted, simulated proof point—not an autonomous field service.

How WFM Twin Works

From owner intent to verified impact.

Stage 1 / Owner Control

The owner defines permission, location, schedule, economics, risk limits, and required supervision.

Proof point 01 / Fuel Ops Loop

One loop. Every decision visible.

Fuel Ops Loop is the current simulated wedge: identify surface-fuel hazards, form cleanup tasks, select a scripted robot, execute the work, and retain the operational trail. It makes the platform thesis concrete without claiming a production wildfire service.

What is real today

An Isaac Sim 6 + ROS 2 Jazzy prototype in Fidei, deterministic task lineage, scripted execution, cleanup events, and an illustrative risk summary.

Event lineage / simulated

01

Observe

Fuel object annotated in scene truth

02

Detect

Hazard and cleanup task created

03

Plan

Robot selected by ROS-side planner

04

Execute

Fuel state changes in simulation

05

Verify

Cleanup event and risk summary emitted

Task horizon

Prove the contract can travel.

The goal is not a list of robot stunts. Each scenario should test whether the same qualification, authorization, assignment, and evidence model survives a different kind of work.

01Working in Fidei

Simulated proof point

Fuel Ops Loop

Observations become hazards, authorized cleanup tasks, robot assignments, execution records, and illustrative risk outcomes.

02Contract reuse test

Candidate scenario

Community Improvement

Explore bounded cleanup work where completion is visible and owner, property, and supervision constraints can be explicit.

03Domain review required

Research hypothesis

Wildlife Guardian

Study observation-first road and habitat workflows without pretending physical deterrence is already safe or validated.

Trust is the product

Dispatch is easy. Governed eligibility is hard.

  • Capability is specific to the robot, tool, task, and environment.
  • Owner authorization is a binding decision—not a note added after dispatch.
  • Simulation evidence is labeled separately from supervised field evidence.
  • Correct refusals are successful outcomes when work is unsafe or unsupported.

Near-term roadmap

Build evidence before scale.

Now

Make the thesis visible

Publish the site, record Fuel Ops Loop, and show what is implemented versus proposed.

Next

Extract the domain contract

Model tasks, capabilities, owner policy, assignments, and evidence outside the simulator.

Then

Earn external evidence

Interview operators and seek one narrowly scoped, supervised design-partner workflow.

An open invitation

Help define the work robots should do next.

I’m looking for robot owners, integrators, operators, wildfire and land-management professionals, insurers, and safety experts willing to challenge the model early.

Talk with Nick