Former DeepMind team launches AI robotics startup with robots that learn from workers

Former DeepMind team launches AI robotics startup with robots that learn from workers

Reimagine Robotics, an AI robotics firm based by former leaders of Google DeepMind’s Utilized Robotics group, has emerged from stealth with new expertise that permits robots to be taught on the job.

The corporate is growing clever robots that anybody can prepare and use. As a substitute of requiring specialist programmers every time a process or manufacturing course of modifications, employees can present the robotic what to do, watch it try the duty and proper it on the spot.

“A helpful robotic ought to be capable to be taught from the individual doing the work,” says Jonathan Scholz, co-founder and CEO of Reimagine Robotics.

“They need to be capable to present it a process, put it proper when it makes a mistake and transfer on to the following downside. That’s what it means for a robotic to be taught on the job. It’s a course of we name ‘monkey-see, monkey-do’.”

“A robotic ought to arrive with the perspective of a brand new colleague: ‘How can I assist? What would you like me to do?’ The individuals who perceive the method ought to be capable to reply these questions by displaying the robotic instantly.”


Scholz based Google DeepMind’s Utilized Robotics group in London and led it for seven years. He later co-founded Reimagine Robotics in April 2025 alongside former colleagues Oleg Sushkov, Akhil Raju and Misha Denil, with headquarters in London and Sydney.

That first section was funded by pre-seed financing from enterprise capital companies Fly Ventures and firstminute capital and quite a few angel traders.

The subsequent section

“We’re excited to come back out of stealth.” Scholz says. “This final 12 months was about constructing a core product and a group, and dealing with clients to check the platform.

“After working with a number of companions, we’re completely satisfied it’s not solely viable, however that there’s a want for this expertise, and that it has large potential throughout an array of commercial and manufacturing settings.

“This subsequent stage is a couple of new spherical of fundraising, increasing our group for extra deployment muscle, placing robots into extra workplaces, and displaying that every deployment could make the following one sooner, extra dependable, and extra environment friendly.”

Robots already on the manufacturing facility flooring

The corporate is already deploying its robots in superior manufacturing and electronics disassembly services.

At a made-to-order plastics enterprise, Reimagine Robotics skilled its robots to have a tendency 3D printers in a single day by eradicating print beds, working latches and urgent controls. The shopper’s personal group used the platform to automate extra levels, together with washing, curing and drying.

In a separate deployment involving the restoration of priceless crucial supplies from used arduous drives, Reimagine Robotics labored with course of engineers to develop a three-robot disassembly cell.

The ensuing workflow combines robots and folks working collectively to evolve and optimise the workflow in real-time.

In the course of the challenge, Reimagine Robotics diminished the time required to prototype and take a look at a brand new robotic behaviour from roughly at some point to round ten minutes.

That allowed the robots to turn into a part of the process-design dialog: groups may suggest a brand new use for a robotic and take a look at the thought virtually instantly.

People on the coronary heart of robotics

All through all Reimagine Robotics deployments, individuals stay central to the mannequin.

“For us, this isn’t about taking individuals out of the method,” Scholz says. “A robotic that learns on the job will depend on individuals. The employee identifies the bottleneck, reveals the robotic learn how to assist and corrects it till it’s helpful.”

“The robotic turns that individual’s information into leverage. As a substitute of somebody having to repeat a tedious bodily process hundreds of occasions, they will educate the robotic, and apply that means wherever it’s wanted.”

“I consider it extra as a software to amplify human labour.”