Food Science Prof Bridges Food Fermentation With Machine Learning to Drive Innovation

Sep 10, 2025

Digital technologies are revolutionizing how food is produced, processed and developed. Tools like advanced biosensors and machine learning are enhancing efficiency, advancing sustainability, improving health outcomes and unlocking the next generation of food ingredients.

Dr. Biniam Kebede, a newly appointed professor in the Department of Food Science at the Ontario Agricultural College, leads the Food Bioprocessing and Data Science Lab, where ancient fermentation meets cutting-edge digital innovation.

“Fermentation is increasingly recognized as a natural and sustainable food processing method,” says Kebede. “But most processes still rely on trial and error. There’s a critical need for data-driven approaches that accelerate R&D and streamline innovation.”

Kebede joined the University of Guelph in December 2024, following a senior lectureship at the University of Otago (New Zealand) and a PhD in bioscience engineering from KU Leuven (Belgium).

His NSERC Discovery Grant-funded research focuses on producing bioactive and flavour compounds from underutilized plant materials (such as pulses) and agri-food byproducts through solid-state fermentation, integrated with multi-omics analysis and AI-driven modelling.

Click here to see more...

Video: Seeing the Whole Season: How Continuous Crop Modeling Is Changing Breeding

Plant breeding has long been shaped by snapshots. A walk through a plot. A single set of notes. A yield check at the end of the season. But crops do not grow in moments. They change every day.

In this conversation, Gary Nijak of AerialPLOT explains how continuous crop modeling is changing the way breeders see, measure, and select plants by capturing growth, stress, and recovery across the entire season, not just at isolated points in time.

Nijak breaks down why point-in-time observations can miss critical performance signals, how repeated, season-long data collection removes the human bottleneck in breeding, and what becomes possible when every plot is treated as a living data set. He also explores how continuous modeling allows breeding programs to move beyond vague descriptors and toward measurable, repeatable insights that connect directly to on-farm outcomes.

This conversation explores:

• What continuous crop modeling is and how it works

• Why traditional field observations fall short over a full growing season

• How scale and repeated measurement change breeding decisions

• What “digital twins” of plots mean for selection and performance

• Why data, not hardware, is driving the next shift in breeding innovation As data-driven breeding moves from research into real-world programs, this discussion offers a clear look at how seeing the whole season is reshaping value for breeders, seed companies, and farmers, and why this may be only the beginning.