Founding Simulation & Risk Modeling Engineer (Robotics)

Robots are leaving pilots and going into production: factories, warehouses, hospitals, hotels, data centres. Insurance hasn't caught up. There's little loss data, and nobody agrees where liability sits between the manufacturer, the integrator and the site operator, so most deployments run underinsured or uninsured.

We're building the insurance layer for robot deployments. We start as a US commercial property-casualty brokerage that assesses deployment risk properly and places cover. Over time we'll become an MGA that underwrites with data no carrier has today.

You'd be one of the first people on the team, so you'd shape the product and the company, and you'd be taking real early-stage risk.

The role

Type: paid contract now, full-time founding role on funding. Comp: $80–160K base on funding (adjusted by location) + 0.5–1% equity. Location: remote, worldwide (US time-zone overlap needed).

Insurers can't price robot risk because they have no loss history. You'll build the substitute: simulated deployments, real telemetry and a model that predicts what goes wrong, how often and how badly.

What you'll do

• Build simulation environments that reproduce real deployment sites and inject failures: sensor faults, collisions, people entering cells, network drops.
• Build an adapter that pulls scenarios in from other simulation stacks, so manufacturers can plug in what they already run.
• Turn partner telemetry and incident logs into risk features.
• Ship v1 of a predictive risk model (frequency and severity) and the evaluation harness behind it.
• Work with our manufacturer partner's engineers on what data we collect and how.

You probably have

• 3+ years in robotics simulation, RL or perception, working with real robot data and not only benchmarks.
• Hands-on Isaac Sim, MuJoCo or Gazebo, plus ROS 2 and Python.
• Something you shipped on your own, end to end.

Nice to have: reliability engineering, survival analysis, safety cases, or actuarial or insurance modelling.

After any interview we'll share specific feedback, whatever the outcome.

To apply: in your note, link one robotics, simulation or risk-modelling project you built yourself (repo, paper, demo or write-up) and say what you did on it. Applications without one won't be reviewed.