A destination · not a claim
Building the physical-intelligence layer for robots.
Machines that can perceive, decide and act reliably in the physical world—from Earth’s hardest worksites to infrastructure beyond it.
The Moon is the first horizon.
Origin · age 13
This story began in a workshop.
I started building competition robots after school, learning mechanics, electronics and code by making the machines real.
30+ national-level winsRoboWar · Robo Soccer · Maze Solver · Mechatronics
Photographic archiveSix machines. Sparks, steel, pneumatics and the first hard lessons in physical intelligence.
Movement · Chisel Robotics
Restore movement.
A dorm-room bionic leg became five years of building for people who had lost a limb—and a lesson in turning research into something physical.
4 NHS trials · under 50 ms feedbackPerformance legs for Paralympic athletes, including a Rio gold medallist.
Production perception
Teach machines to see.
Roads, vehicles and factories forced perception to survive months of real operation—not one good demo.
RoadsideDetection, tracking, speed and traffic flow on NVIDIA edge hardware.
Vehicle inspectionApproximately four hours to under 90 minutes.
Factory scaleHundreds of stations across more than 60 sites.
From perception to action
Give perception a body.
I built a perception-to-action stack for industrial manipulation: learning from demonstrations where it helps, classical control where physics demands guarantees.
Completed customer pilotFranka Research 3 Duo · Universal Robots · contact-rich tasksGoogle DeepMind Robotics Accelerator · MassRobotics Physical AI Fellowship
The operating principle
Earth is the proving ground.
Real customers, strict safety requirements and immediate consequences are how dependable autonomy gets built—for factories, heavy machines and difficult environments.
Perceive → understand → plan → actThe same loop connects roads, robot arms and heavy equipment.
The long game
Build where people cannot yet safely live.
Before people can remain beyond Earth, machines will need to prepare landing areas, move material, assemble habitats and keep infrastructure working with limited supervision.
I am not building lunar machinery today. I am deliberately building the perception, manipulation and autonomy such machines will eventually need.
Explore the full journeyThe path is long.
The direction is clear.
