We build
spatial intelligence
The last stretch of automation is the coordinate on the floor.
The robots existed and the work existed.
The thing joining them did not.
We make the coordinates
the floor runs on
A robot cannot work until it knows the state the plant is in right now. We read the whole plant in 3D, build the coordinates a robot can act on, and take over the work people were doing on top of them.
No process should stop
for want of people
Shipbuilding and heavy industry are running short of people while automation has not reached them. We are turning the stretch that could never be written as rules into work a robot can take.
How we work
We gather it in real yards, not in simulation. Delivering a project and building the ground for the next task happen at the same time.
Ten photographs to adapt to a site we have never seen, and not a millimetre of drift once the work starts. Flexibility and precision belong in different layers.
A cheaper or better robot arrives and we swap it in. What we make is not the robot but the coordinates the robot uses. Plenty of companies build sensors, robots and task policies. The layer that reads a whole plant and supplies those coordinates was empty.
We do not take on projects that end at a demonstration. A job is done when work people used to do has actually been handed over.
Outside the line
Shipbuilding and heavy industry are that outside. Every part has a different shape and lands in a different place, so the work of looking and judging has stayed with people.
It was left behind because it could not be written as a rule, and that is exactly why it is priced highest.
LiOps works outside the line.
Who builds it
Perception and control are built by one team.
When the vision company and the robot company are separate, a drift in coordinates turns into an argument and a person ends up teaching the path again. So a single team holds the whole chain, from sensor to robot.
We spend more time in the yard than in the office, because we do not buy data, we make it on site. Everything from scanning to verification happens inside the shipyard.
- 3D perception for autonomous driving · Hyundai Mobis, LiDAR research at VueRon
- On the record · First worldwide on the Waymo 3D perception leaderboard
- 3D vision research · CVPR main-track authors
- Autonomous robots · AMR engineers from Yujin Robot and Polaris3D
- Robotics · Biorobotics Lab, Seoul National University
- Software · MLOps at RECON Labs, ten years of full-stack work
How we got here
Many places, one goal
Seongsu SK V1 Tower
Godeok G1 Knowledge Industry Center
We are looking for people
to build LiOps with
Hardest site, hardest task, starting there.