How a Smart Agriculture App Built by Students and Powered by JamAI Base is Reimagining The Future of Food Security.
At the 2025 FoSEAL Hackathon, hosted by Universiti Putra Malaysia (UPM), a team of innovative students from the University of Science, Malaysia dug deep into the global challenge of food security and emerged victorious. Their solution, GrowSmart, is a precision agriculture platform that blends AI and IoT to help farmers optimize resources and boost yields. Powered by JamAI Base, this app is designed to be a practical, farmer-friendly tool that makes a real-world impact.
Refined from a previous hackathon project, the team took judges’ feedback to heart, enhancing GrowSmart to stand out in UPM’s competitive sprint. Their hard work paid off, earning praise for its innovative features and ambitious vision.
What They Built: GrowSmart in Action
Source: Presentation deck of GrowSmart
GrowSmart is a digital co-pilot for farmers, addressing key challenges with smart, scalable features:
- Smart Irrigation & Fertilization: AI-driven schedules deliver water and nutrients precisely, reducing waste and environmental impact.
- Real-Time Monitoring: IoT sensors track soil moisture, pH, temperature, and humidity, while satellite data flags weather risks early.
- Precision Farming & Decision Support: GPS zoning and AI yield forecasts optimize field management, with real-time alerts for soil or crop issues.
- AI-Powered Crop Health Analysis: Image recognition detects pests, diseases, and deficiencies, paired with a chatbot offering instant advice in a conversational tone.
The judges were particularly impressed with GrowSmart’s multilingual support plan, tailored for Malaysia’s diverse farmers who often mix languages like Hakka, Cantonese, English, and Malay. This feature ensures accessibility for farmers with varying education levels.
Behind the Tech: How JamAI Base Made It Possible
GrowSmart’s intelligence runs on JamAI Base, an open-source, enterprise-grade AI platform that empowered the team to build a robust solution within a week. Key features of JamAI Base that drove their success include:
- Low-code Development with Action Tables: Instead of wrestling with complex node graphs, the team used JamAI Base’s spreadsheet-like Action Tables. This kept their data and the AI’s results visible side-by-side, creating a rapid feedback loop that saved critical time and allowed them to perfect the AI’s logic in hours, not days.
- Instant Chat Interface with JamAI Chat: To create a polished demo for the judges, the team instantly deployed their AI logic to JamAI Chat. This gave them a farmer-friendly, conversational interface without writing a single line of front-end code, allowing them to focus entirely on the AI’s intelligence.
- Transparent AI Logic for Domain Experts: The spreadsheet-like interface makes the AI’s “thinking” transparent. An agronomist can open the Action Table, see a photo of a diseased leaf in one column and the AI’s diagnosis in the next, and instantly validate or correct the logic. This builds trust and ensures the AI is grounded in real-world agricultural science.
- Portable AI Logic with Generative Tables: The team packaged their entire AI workflow into a single, portable “Generative Table” file. This self-contained “brain” holds all the AI input, output, AI logic chain and prompts, making their prototype robust and easy to manage under pressure. It was their secret weapon for creating a professional-grade solution.
- Simplified Future Updates for Experts: The portable “Generative Table” means an agronomist can later refine the crop disease rules or add new ones by simply editing the table. Deploying this smarter AI is as easy as swapping a file, requiring no code changes from the development team and ensuring the app stays up-to-date with the latest science.
The team initially considered mainstream AI tools like Gemini or OpenAI, used by many competitors, but chose JamAI Base after a workshop introduction by the JamAI team. Its unique features and dedicated support made GrowSmart stand out, with the team celebrating the moment their code ran error-free after nights of debugging.
Source: Presentation deck of GrowSmart
Overcoming Challenges
The hackathon’s tight timeline was the team’s biggest hurdle. With only a few days to brainstorm and build a functional prototype, they faced intense pressure. As university students, they also juggled final exams, assignments, and other events during the grand finals, even rescheduling a test to compete. Despite these constraints, their focus on a minimum viable product and rapid iteration with JamAI Base secured their spot in the finals and a polished solution by the end.
Looking Ahead
The team sees GrowSmart as more than a hackathon win — it’s a game-changer for Malaysia’s agriculture industry. By combining AI with local and internet-sourced data, they aim to modernize farming practices that often rely on outdated experience. Their vision includes field tests with local farmers, adding market trend analytics, and securing funding to bring GrowSmart to market. With JamAI Base’s scalability, they’re confident in its potential to optimize crop yields and make a lasting impact.
We’re thrilled to see JamAI Base fueling solutions that tackle global challenges like food security. GrowSmart proves that when open-source AI meets passionate innovators, the results can transform industries. Want to build something world-changing? Check out JamAI Base on GitHub or dive into our quickstart tutorials. The future of agriculture is here — let’s grow it together!
#FoSEALHackathon2025 #AIforGood #FoodSecurity #PrecisionFarming #OpenSourceAI
Source: Virtual interview with the team
