This London startup wants factory workers to train robots the same way they would train a new hire

Key points

• Reimagine Robotics, a London and Sydney startup, came out of stealth in 2025 after roughly a year of quiet development and customer testing.
• The company's technology lets ordinary workers teach robots new tasks by demonstrating them directly, then correcting mistakes on the spot, with no coding required.
• In a hard-drive disassembly project, Reimagine cut the time to prototype a new robot behaviour from about one day to roughly ten minutes.
• The startup's founding team includes CEO Jonathan Scholz, who built and ran Google DeepMind's Applied Robotics team in London for seven years.
• Reimagine has already deployed robots in a 3D-printing plastics business and an electronics recycling facility, with more deployments and a new funding round planned.

Teaching a robot used to mean hiring a specialist programmer and waiting days. Reimagine Robotics, a startup that came out of hiding this week, says it has a different answer: let the person who actually does the job do the teaching instead.

The company's system works on a show-and-correct principle. A worker demonstrates a task, the robot tries to copy it, and the worker steps in to fix any mistakes. No code. No robotics degree required.

"A useful robot should be able to learn from the person doing the work," said co-founder and CEO Jonathan Scholz, who previously founded and led Google DeepMind's Applied Robotics team in London for seven years before co-founding Reimagine in April 2025 with former colleagues Oleg Sushkov, Akhil Raju, and Misha Denil.

What does it actually do in practice?

The company has already put robots to work in two real facilities, not just a lab. At a made-to-order plastics company, Reimagine's robots learned to tend 3D printers overnight, removing finished print beds, operating latches, and pressing controls. The customer's own team then used the platform to add extra stages: washing, curing, and drying parts.

A second project involved recovering valuable materials from used hard drives. Reimagine worked with process engineers to build a three-robot disassembly cell that combined machines and people working side by side. During that project, the time needed to test a brand-new robot behaviour dropped from roughly one day to about ten minutes, according to the company. That meant engineers could come up with an idea in the morning and see it running before lunch.

Deployment	Task taught	Key result
Plastics manufacturer	3D-printer tending, washing, curing, drying	Workers extended automation themselves, no programmer needed
Electronics recycler	Hard-drive disassembly, three-robot cell	Behaviour prototyping time: ~1 day to ~10 minutes

Scholz is careful to frame the technology as a tool that sits alongside people, not one that replaces them. "Instead of someone having to repeat a tedious physical task thousands of times, they can teach the robot, and apply that ability wherever it is needed," he said.

What happens next?

A new funding round is the immediate priority. The startup's first phase was backed by pre-seed money from Fly Ventures, firstminute capital, and a group of angel investors. Scholz says the next stage will focus on hiring more deployment staff, expanding into additional factories, and proving that each new installation gets faster and cheaper than the last.

As first reported by The Robot Report, Reimagine joins a crowded field of companies trying to make robots easier to train, including 1X Technologies, Apptronik, and Sanctuary AI. The differentiator Reimagine is betting on is speed: the faster a factory worker can teach a robot something new, the more useful it becomes as products and processes change.

For workers on the factory floor, the practical message is straightforward. If this technology delivers on its promise, training a robot may one day look less like writing software and more like onboarding a new colleague.