Two days at Manufacturing Forum 2026 in Mladá Boleslav
On 11 and 12 November 2026, this year's Manufacturing Forum took place at the Škoda Museum in Mladá Boleslav. We brought a stand with a live camera inspection demo, and Jakub Šnorbert gave a stage talk on how we check baguette quality directly on the line at Crocodille. Read on to see what manufacturing visitors asked about most, and watch the recording of the talk on YouTube.
Checking thousands of baguettes an hour to make sure not a single one is missing its egg? Artificial intelligence can handle that today. Yet as Jakub Šnorbert, Head of Delivery at Robopipe, showed at Manufacturing Forum 2026 at the Škoda Museum in Mladá Boleslav, the real value of technology like this goes much further than first meets the eye.
In his talk on Wednesday 11 November, he shared the story of our work with baguette producer Crocodille, whose lines turn out up to 3,600 baguettes an hour. Every single baguette has to pass a quality check that staff previously carried out by hand.
The project started with a straightforward goal: use artificial intelligence to spot baguettes missing their egg. Robopipe installed a camera on the production line equipped with a specially trained neural network. In real time, it recognises individual ingredients and checks both their quantity and their placement.
What began as a simple check for eggs gradually grew into a broader inspection system that now runs 19 checks on every baguette. Crocodille's own team defined these criteria around their needs on the line, so the system flags not only missing ingredients, but also ingredients sitting in the wrong place.
The technical setup is worth noting too. Because the neural network runs directly inside the camera, it works without an internet connection. Inspection runs continuously on the production line, and Crocodille's staff can follow the results on a clear dashboard.
The bigger benefit
The greatest payoff, however, only became clear over time. Using data from the cameras, Crocodille began comparing individual shifts, baguette varieties and production lines in detail. That allowed them to uncover not just defects themselves, but their root causes. Often, the real issue is not one baguette missing an egg, but a flaw somewhere in the production process.
That is where automated quality control turns into a tool for optimising production. The Robopipe dashboard highlights emerging issues in real time right beside the line, so operators can step in before a minor hiccup spoils the output of an entire shift.
Nor is the technology limited to food manufacturing. As Šnorbert explained on stage, Robopipe can just as easily check item counts, packaging, batch codes or vehicle number plates. Each customer sets their own inspection criteria, and the neural network is trained on data from their own operation so it reflects local conditions and delivers accurate results.
Live demo
Visitors could also see the system at work in person. At the Robopipe stand, a camera mounted above a worktop inspected a metal component in real time. While a tablet displayed the result of each check, a large screen beside it ran the live dashboard with collected data, showing how automated object recognition links directly to instant evaluation.
Our stand was right in the same hall as the stage. Plenty of attendees stopped by the demo over the two days, and we had in-depth conversations with 25 to 35 of them. Most came from manufacturing, particularly automotive companies, production line builders and component makers. Their questions went straight to practical details: which model we train our neural network on, whether we can measure distances as well, and which camera we use.
What comes next
Those conversations at the stand brought us several interesting problems that will be a welcome challenge for our developers too. Over the coming weeks, we will be heading out to visit manufacturing plants in person.
The event itself was a pleasant surprise, above all because of who turned up in Mladá Boleslav. We spoke directly with the people who decide on technology investments inside their companies. Our thanks go to the organisers of Manufacturing Forum, and to everyone who came to the talk or stopped by our stand for a chat.
Video of the talk
You can watch the recording of Jakub's talk, What do eggs in a baguette and AI have in common?, on YouTube (in Czech). For a closer look at how the system works on the line, read our Crocodille case study.
Get in touch
If we missed each other in Mladá Boleslav and you are tackling a similar problem in production, get in touch through our contact page. We would be glad to discuss quality control on your line and arrange a live demo at your plant.