About Eyecan
Eyecan was founded in 2020 as a spin-off from the University of Bologna, with the initial mission of commercializing patented technology (dubbed "The Training Machine") designed to train robotic picking applications without manual annotations.
In 2021, Marchesini Group S.p.A. joined the company in a seed round to validate the product.
In 2026 (today), Marchesini and other major industrial leaders deploy the platform in production—with over 200 AI Models deployed worldwide powered by Eyecan,
Where it all began...
Our very first prototype of "Training Machine", engineered to generate infinite training data straight from real-world physical objects scans.
Our Team & Story
Eyecan was born in the Computer Vision Lab at the University of Bologna, led by Prof. Luigi Di Stefano, sparked by a radical idea from founder Phd. Daniele De Gregorio while studying under Prof. Di Stefano.
The breakthrough behind Eyecan came from a stroke of serendipity. Instead of using computer vision to guide robotics as usual, Eyecan flipped the logic: use robotics to train vision.
This led to a patented technology where robots generate the data needed to train other robots, pioneering the industry's first approach to training AI without manual annotation.
The patented technology worked brilliantly. But breaking into the industrial market was a whole new challenge...
Eyecan’s underlying tech was rock-solid, powering custom AI models across our clients' toughest applications.
But software alone wasn't the answer. We hit a scaling bottleneck: operating the system required specialized computer vision expertise that standard industrial teams lacked, and we couldn't act as a permanent R&D team for every customer.
Enter our "dogfooding" breakthrough. We took the internal SaaS platform our own team used to deploy AI and converted it into a client-facing web app.
We replaced complex R&D overhead with an intuitive, no-code platform built for both humans and AI agents.
Now, manufacturers and machine builders are fully independent. Complex AI runs quietly under the hood, delivering industrial-grade robustness right out of the box.
