News

Jun 10 '25

From Idea to Impact: Students Use AI x Sony Tech to Solve Food System Challenges

#AIFS

#Education

Steve Brown introducing students to the Sony Spresense
An image without an alt, whoops
An image without an alt, whoops
An image without an alt, whoops
An image without an alt, whoops
An image without an alt, whoops
Previous
Next

Empowering Future Innovators Through Hands-On AgTech Challenges

Over one action-packed weekend, five teams of students channeled their curiosity and technical skills into practical solutions for agriculture and food systems. The AIFS x Sony AI AgTech Challenge, held May 2–4, 2025, at UC Davis, brought together undergraduate and graduate students from a range of disciplines to prototype tools using Sony Spresense microcontroller kits and Raspberry Pi-connected Sony AI Cameras. Their projects combined AI, hardware, and hands-on problem-solving to develop technologies designed for real-world agricultural applications.

Event Overview: Building and Pitching Tech for Real-World Agriculture

The weekend kicked off with an engaging talk from Dr. Emmanouil (Manos) Koukoumidis, CEO of Oumi PBC. He spoke about the need for open-source frontier AI — AI that grows through shared effort, builds trust, and serves the public good. Students were invited to contribute to Oumi’s open GitHub project.

Following the keynote, AIFS Associate Director Steve Brown provided an overview of the challenge, outlining project goals, available resources, and judging criteria. Team assignments were finalized, and participants engaged in a Q&A session before beginning development.

Each team combined skills from engineering, computer science, and food science to tackle a pressing challenge in the food system. After just 48 hours of building and testing, teams pitched their projects to a panel of judges. In a rare outcome, two teams, RipeRight and WeedTrakr, tied for first place.

The event highlighted how interdisciplinary training and industry collaboration can accelerate the development of practical AgTech solutions.

Team Spotlight: AgChat

The AgChat team created a smart fertilizer assistant for strawberry farmers, especially those in under-resourced regions. This approach blends a planning model with real-time leaf analysis, helping farmers make timely, data-informed decisions.

By using a Raspberry Pi-connected AI Camera and a custom-trained GPT-4 model, the tool detects nutrient issues and adjusts fertilizer schedules. This "virtual agronomist" reduces reliance on on-site consultants and offers small farms a more cost-effective solution.

Team Spotlight: AutoDrive

AutoDrive developed a small autonomous robot to help farmers estimate yields and monitor crops. Built on an RC car with a Sony Spresense kit, it drives through rows of plants, uses AI to count fruit, and logs the data along with GPS coordinates.

This approach reduces guesswork and supports more accurate planning. It could help farmers better manage labor, plan logistics, and forecast income with less effort and more confidence.

Team Spotlight: YOLOsort

Sorting fruit by hand is slow and inconsistent. The YOLOsort team set out to change that. They trained an AI model to identify grapes and cherries in real time and built a working prototype using Raspberry Pi components.

Their system speeds up sorting, improves hygiene, and enables data tracking for traceability. With further work, it could sort by size or quality and plug into existing packing lines.

Team Spotlight: WeedTrakr

WeedTrakr aims to help farmers target weeds precisely, cutting down on broad chemical use. Their tool, mounted on a tractor or robot, uses AI to spot weeds and GPS to map their locations.

By pinpointing problem areas, WeedTrakr can help farmers reduce herbicide use, lower costs, and manage fields more sustainably. This approach benefits both environmental sustainability and farm efficiency.

Team Spotlight: RipeRight

Knowing exactly when fruit is ripe can make or break a harvest. RipeRight built a tool that uses image analysis to detect peak ripeness in strawberries. Their prototype analyzes color and texture cues using a Raspberry Pi setup and a custom dataset.

Better timing means less waste and better-tasting produce. The project offers an accessible tech option for farms of many sizes.

Celebrating Talent and Collaboration

Faculty lead Dr. Zhaodan Kong reflected the strength of the program and student outcomes: "It was an absolute success! I was genuinely amazed by how seamlessly everything went and deeply inspired by the outstanding quality of the students' work."

Student feedback echoed that praise. Charlene Hui wrote, "It helped me to think about what I can do with AI and technology even without a computer science or engineering background." Jonny Berlingeri commented, "The Challenge was a fun collaborative coding and hardware experience. It was exciting and felt like what I would imagine the beginning of a start up might feel like."

Looking Ahead: From Campus to Classroom

The event also inspired future program development, including a summer camp for high schoolers using the Sony components. Led by Dr. Kong, students will replicate the projects and learn how each one works, offering an early glimpse into tech-driven agriculture.

Acknowledgments

Thank you to Dr. Koukoumidis for his keynote, Dr. Kong for his leadership, Sony for its generous support, and the many mentors and judges who volunteered their time. Most of all, thank you to the students for bringing passion and ingenuity to every project.

Congratulations again to RipeRight and WeedTrakr for sharing first place — and to all five teams for their thoughtful, hopeful, and forward-looking contributions to the future of food.

Related News & Events