Mar 10, 2025
A 36-hour hackathon at Cornell led our team from a small, three-person group with no concrete idea to a five-person team building a case study, a business model, and a working prototype for a highly specific product in a disaggregated market.
Day 1: Initial Idea
Our team formed about a week before the hackathon began. We knew the theme was Digital Agriculture, but we weren’t sure what to build. So, the research started by reading the tracks specified for the hackathon which were as follows:
Since our team was made up of data and computer scientists, we gravitated toward the Data-Driven, Human-Centric Digital Agriculture Innovations track, aiming to build an AI solution. Through our research, we discovered that food wastage is one of the biggest issues in agriculture:
Roughly one-third of all food produced — around 1.3 billion tonnes — is either lost or wasted every year (Food and Agriculture Organization)
This loss is estimated to translate to economic damages of around $1 trillion each year
Food wastage, however, is an extremely broad topic — applicable to farms, restaurants, consumers, and retailers. We decided to focus on the crop flow process, specifically on potatoes. Why potatoes? For one simple reason, who doesn’t love fries (yummmm) plus, our research indicated that 30% of potato yield is lost in the post-harvest stage at the trader level.
This image shows the Crop Flow process for potatoes from Farmer to Consumer.
Just as the hackathon started, we recruited two new members. Our goal was to build a dashboard to raise awareness about the post-harvest stage and to promote novel storage strategies, reducing grading failures and increasing overall production. Our target users were farmers, wholesalers, and retailers.
Day 1 had ended.
Day 2: Mentorships and Pivot
Day 2 began with industry leaders visiting our workspace to hear about our solution and provide insights. We pitched our ideas, and they gave us feedback on feasibility, improvements, and potential pitfalls. We felt confident until the next mentor asked some hard-hitting questions:
“Why potatoes? What are you solving?” — Just because everyone likes fries isn’t a good enough reason. Yes, there is food wastage but that’s for all crops.
“Which location are we targeting and why?” — each location has different weather conditions, different soil qualities.
“Who are we targeting exactly? What is the farmer or retailer profile?” — for every successful product, there is a highly targeted user base.
These questions forced us to reevaluate our solution. We realized we needed a clearer focus.
We gained a lot of clarity on what we were lacking.
Crops can be categorized by perishability: low, high, and extremely perishable. The more perishable, the more complex the problems — and the more opportunities to make an impact.
Crop loss can occur at each stage of production.
The mentors helped us identify a more targeted problem:
highly perishable items + disaggregated markets
After further discussion, we zeroed in on tomatoes.
Here’s why:
Most farms are small scale or family operated
Production is spread out and not coordinated
Lots of moving parts in the harvest to consumer space
Diverse range of producers, quality, size — challenging to establish uniform quality standards
We chose India as our launch market, given its massive tomato production and alignment with these factors. Our next step was to refine our idea and build both a product and a business model.
Our Solution: Seed Smart
Through extensive research, we created Seed Smart:
An app that alerts farmers if a particular crop — initially tomatoes — might be overproduced in their area, prompting them to choose alternative crops before the planting season begins.
The app targets small-scale farmers and will start with one state in India.
How It Works
The app maps which crops local farmers have planted.
A threshold is set to avoid oversupply. If tomatoes exceed this threshold, the farmer is prompted to plant another crop.
Credits: @Neha Kulkarni and @Anurag Choubey
Farmers can input their state, district, chosen crop, and acreage.
Data is collected nightly and processed to calculate potential profit or loss using a formula we developed.
Results are displayed on a dashboard and in the mobile app.
Business Model & Timeline
The business model comes with a timeline on where we want the progression to be.
0–3 Months: Secure a pilot and establish partnerships with local stakeholders.
3–6 Months: Drive farmer adoption and expand our data collection efforts.
6 — 12 Months: Focus on securing funding and scaling our solution.
12+ Months: Move toward monetization and broader expansion, ensuring the platform’s sustainability and continued impact.
As Day 2 ended and there were less than 6 hours to go, we had a completely functioning prototype, a pipeline to ingest data and a consumer facing dashboard. The design process must be given some credit as well, since this was built in less than 30 hours.
Figma was a life saver
Day 3: Presentations and the End
Along with other teams, we recorded our demo on Day 3 and concentrated on improving our presentation. The judges offered knowledge on related projects in various areas and provided insightful criticism on how to make our product better.
All things considered, the hackathon was an amazing event with unforgettable moments and educational chances. We came away with new knowledge, new connections, and the satisfaction that comes from creating something that could lessen food waste and help farmers make a living.
Conclusion
The Cornell Digital Agriculture Hackathon challenged us to think deeply about real-world agricultural problems, pushing us to create Seed Smart, a solution that leverages data to reduce oversupply and post-harvest losses.
We look forward to taking this idea beyond the hackathon stage, working with stakeholders, and making a tangible impact on sustainable agriculture.
View Link for the Pitch Deck -
