Russia Brews Beer From an Artificial Intelligence–Generated Recipe
In Voronezh, Russian researchers brewed an experimental beer using a recipe generated by their own neural network, illustrating how AI can reshape consumer‑driven food and beverage production
Students at the Voronezh State University of Engineering Technologies (VSUET) have brewed an experimental batch of beer based entirely on a recipe generated by their custom‑built neural network. The project, developed by graduate students Daniil Kuligin and Vladislav Vasechkin, involved training an algorithm on the flavor preferences of Russian consumers and the characteristics of various beer styles to create a balanced and data‑driven formula.
Barley and Malt
The neural network recommended using pale barley and caramel malt, pellet hops with an alpha‑acid content of 5–6 percent, and aroma hops of the Cascade, Centennial, and Amarillo varieties to create a fruity, expressive bouquet. During fermentation, the students used top‑fermenting Wyeast 1056 yeast and soft water.
The algorithm calculated a wort original gravity of 12 percent, bitterness of 45 IBU, caloric value of 48 kcal, and an alcohol content of 4.18 percent.
Brewed to Industry Standards
The beer was produced under laboratory conditions at VSUET’s Department of Fermentation and Sugar Production. A technologist from a commercial brewery confirmed that the resulting beverage meets industry quality standards.
The developers note that their algorithm can be scaled for industrial use, with built‑in support for integration into enterprise ERP and SCADA systems.
A Craft‑Minded Neural Network
The team sees their work not just as a scientific experiment but as a step toward consumer‑tailored brewing. Over time, the neural network could track shifts in consumer preferences and generate recipes suited to specific demographic groups — or even to an individual’s unique taste profile.
