Ground truth first
We work from real telemetry and real defects, not slideware. The plant floor is the arbiter.
Careers
Physical AI, where the loop closes on real machines
We build software that senses, predicts and acts on corrugators, presses, die-cutters, folder-gluers and robots. If you want your models to move metal, and margin, this is the place.
The stance
How we work
A small team close to the plant floor.
Trust before autonomy
We ship grounded, bounded, auditable systems, and earn the right to close the loop.
Compounding craft
Every correction teaches the system. We build for the flywheel, not the demo.
Own the outcome
We measure ourselves on waste, makeready, claims and OEE, the customer’s numbers.
Open roles
Where we’re hiring
Don’t see your exact role? Reach out anyway.
Engineering→
Engineering→
Robotics→
Simulation→
Field→
Go-to-market→
Perception ML Engineer (Vision)
Fine-tune high-speed print-defect, registration, color, warp and fold-quality models on TensorRT and Holoscan at the plant edge.
Controls & Autonomy Engineer
Design bounded closed-loop setpoint policies and PLC/OPC-UA integration with hard ranges and quality holds.
Robotics Engineer (Isaac)
Build learned grasp/place policies and sim-first cells for feeding, stacking, bundling and palletizing.
Digital Twin Engineer (Omniverse)
Develop the physics-informed converting-line-and-box twin and makeready auto-tuning.
Forward-Deployed Field Engineer
Run pilots on plant floors: baseline, shadow, assist and autonomous, and turn drift into outcomes.
Solutions Engineer (Converting)
Translate corrugator, press and finishing pain into scoped pilots and ROI models.
Why it matters
The craft is retiring. Help us keep it on the floor.
The veteran operators who can feel the drift are scarce and aging. The work here is to encode that craft into systems that hold the standard on every shift, in every plant.
- Real physical-AI loops on real machines
- A cross-machine dataset that compounds
- NVIDIA edge-to-twin stack, end to end
- Outcomes you can measure in margin
EarlyTeam stage
3–5Design partners
Edge → TwinStack
Every boxImpact
What we offer
The deal
Meaningful ownership
Early equity and real scope. You will own systems that ship to plants, not tickets.
Physical-AI depth
Work across perception, control, robotics and simulation on the full NVIDIA stack.
Close to the customer
Regular time on plant floors. The feedback loop is measured in shifts, not quarters.
Mission that lasts
Build the operations layer for the world’s packaging converting.
The best day here is watching a model you trained hold a corrugator to strength through a humidity swing that used to cost a shift of waste.
Answers
Questions converters ask
Do I need packaging experience?
No. We need strong ML, controls, robotics, simulation and field engineers. We’ll teach you the converting physics, and you’ll learn fast on the floor.
Is this remote?
Engineering can be largely remote, but everyone spends time on plant floors. Field roles are on-site with customers.
How do I apply?
Reach out via the contact page with the role and a note on the hardest physical-AI or controls problem you’ve shipped.
Come move metal and margin
Tell us what you want to build. We’ll tell you where it fits on the plant floor.