Software Engineer, Robotics

CircuitHub, Remote +2

What you will do

Architect the robot operating system. Design the core abstractions: how robots, cameras, tools and jobs are modelled, how work is planned and dispatched, how state is shared and how failures are handled. Decide what belongs in Python and what needs Rust.
Bring in new vision. Integrate novel computer vision algorithms into the control loop, from finding and centring parts to checking every pick and placement, and make them fast and reliable enough for production.
Build agentic robotics. Put large models to work in the robots themselves: programming a new part from its datasheet, judging a pick from a camera image, diagnosing a fault and recovering from it, with evals that prove it works before it runs alone.
Move a fleet. Route dozens of movers at once without collisions or deadlocks, and coordinate gantries, printers and ovens around them.
Bridge to real time. Design clean interfaces between high-level planning and the real-time controllers driving the motors.
Make it observable. Simulation, tracing and replay, so every change is proven before it reaches the factory and every failure can be explained.

CircuitHub, Remote +2

Architect the robot operating system. Design the core abstractions: how robots, cameras, tools and jobs are modelled, how work is planned and dispatched, how state is shared and how failures are handled. Decide what belongs in Python and what needs Rust.
Bring in new vision. Integrate novel computer vision algorithms into the control loop, from finding and centring parts to checking every pick and placement, and make them fast and reliable enough for production.
Build agentic robotics. Put large models to work in the robots themselves: programming a new part from its datasheet, judging a pick from a camera image, diagnosing a fault and recovering from it, with evals that prove it works before it runs alone.
Move a fleet. Route dozens of movers at once without collisions or deadlocks, and coordinate gantries, printers and ovens around them.
Bridge to real time. Design clean interfaces between high-level planning and the real-time controllers driving the motors.
Make it observable. Simulation, tracing and replay, so every change is proven before it reaches the factory and every failure can be explained.