From Rice Fields To Accelerating Vegetable Breeding

By combining artificial intelligence, advanced genetics and computer simulations, NC State University doctoral student Kashish Grover is giving plant breeders back their most precious commodity: time.

In the Department of Horticultural Science, working alongside faculty advisor Roberto Fritsche-Neto, an associate professor of horticulture and principal investigator for the Vegetable Prediction Breeding Lab, Grover develops high-tech tools designed to streamline the way new, climate-resilient crop varieties reach growers. His projects range from automated camera systems that evaluate cucumbers in seconds to predictive computer models.

Traditional plant breeding relies on labor-intensive field trials and manual evaluations that can span over a decade. Grover’s research blends conventional breeding knowledge with genomic selection, computer vision and digital simulations, allowing scientists to test breeding strategies virtually before ever planting a seed.

“The challenge I care most about is time,” says Grover, who earned his bachelor’s degree in agriculture and master’s degree in plant breeding and genetics from Punjab Agricultural University in India. “Developing a new variety takes years, much of it spent measuring plants by hand and testing ideas one season at a time. A changing climate adds another layer, because the places we test today may not reflect where crops will grow tomorrow. My projects try to give breeders some of that time back.”

And his passion and dedication have yielded many meaningful contributions.

“Grover has an extraordinary ability to learn new things and develop many projects in parallel,” Fritsche-Neto says. “From greenhouse to computer models, he has helped us to develop new varieties and tools.”

Bridging Genetics and Computer Science
At the center of Grover’s current work is an automated, image-based phenotyping system designed for cucumbers. Manually evaluating traits such as fruit length, shape, color, and internal structure by hand is a grueling process when researchers must assess thousands of samples.

Assisted by visiting scholars Hercules dos Santos and Felipe Roberto Francisco, Grover developed a system that automates this evaluation through digital photography and image analysis, a technical leap that required bridging plant genetics with computer science.

“Teaching a camera to see a cucumber the way a breeder does was harder than I expected,” Grover says. “Color shifts with the lighting and traits a breeder judges at a glance must become precise, repeatable measurements.”

In collaboration with the Ornamental Plant Breeding and Cytogenetics Lab, led by Hsuan Chen, assistant professor of horticultural science, Grover is developing tetraploid cucumbers, plants featuring four sets of chromosomes rather than the standard two. Crossing these with regular cucumbers creates seedless triploid hybrids, applying the same cytogenetic principle behind seedless watermelons, grapes and bananas.

“This ploidy-manipulation breeding strategy has some basis in cucumber breeding; however, making it feasible and efficient is critical for breeding commercial triploid cultivars,” Chen says. “Grover’s project focuses on making this process much more efficient so that breeders can generate more tetraploid germplasm. His research identified the optimal procedure for developing large numbers of tetraploid lines so breeders can find the best hybrid combinations in later breeding steps. Grover’s research is carrying seedless cucumbers one step closer to our table.”

Broadening the Scope
Grover’s predictive research extends well beyond fruit morphology.

He recently published the study “Optimizing Pickling Cucumber (Cucumis sativus L.) Multi-Environment Trial Networks Using Envirotyping, Production Importance and Future Climate Projections,” identifying optimal locations for pickling cucumber breeding trials under current and projected future climate scenarios. In tandem with industry partners, he is also leading soybean simulation studies that compare how distinct breeding strategies impact genetic progress, diversity and operating costs.

“Simulations are my favorite example,” Grover explains. “Instead of spending years testing every strategy in the field, we can compare them on a computer first and take only the most promising ones outside. Breeders never have unlimited time or money and I love helping figure out where those resources are best spent.”

From Roots to Research
Grover’s pathway to agricultural data science began in his grandparents’ village in Haryana, India, where evening walks alongside local rice paddies sparked an early fascination with crop science.

“I spent my childhood summers there and every evening we walked along the rice fields,” Grover says. “I loved those walks, wondering how the small plants around me would one day become the rice on our plates.”

After finishing his master’s research on barley in India, Grover moved to the United States in January 2024 to begin his doctoral studies with Fritsche-Neto at Louisiana State University, where he was studying genomic prediction in rice. A year later, he transferred to NC State, following Fritsche-Neto and becoming a founding member of his new lab within NC State’s College of Agriculture and Life Sciences. Building the lab from the ground up required Grover to manage greenhouse operations, adapt to a new crop category and work across disciplines with collaborators from Michigan State University.

Expected to graduate in 2027, Grover remains active in campus leadership as treasurer of the Horticultural Science Graduate Student Association and as an NC State Small Pack Leader mentoring international students. Supported by the National Association of Plant Breeders, several science society membership networks and NC State endowments, like the Charles W. Stuber Graduate Student Award Endowment, John D. Leazar Horticultural Food Crop Production and Management Endowment and the Dale C. Bone and Genia Tyson Bone Specialty Crops Program Endowment, he frequently presents his findings at national scientific conferences.

Looking ahead to a career in research and development, Grover plans to bring a “founder’s mindset” to agricultural technology, using artificial intelligence and predictive modeling to solve complex breeding challenges.

“Say yes to new opportunities, even when you don’t know where they’ll lead,” Grover advises prospective graduate researchers. “You’ll learn most of what you need along the way, so don’t worry about having everything figured out on day one.”