Low-cost, automated coffee roaster for small-scale farmers
Automated, AI-powered, low-cost coffee roaster for farmers in rural Latin America
Pitch us on your solution
Problem: Farmers across the world rely on coffee for their livelihoods. In Colombia, more than 300,00 thousand families grow coffee as their main economic activity. However, many of these farmers are unable to earn a reliable living from the coffee they grow. Today's coffee value chain rewards farmers capable of selling specialty coffee, meaning the beans have been grown using special techniques, and/or roasted. Although farmers have access to growing techniques, roasting is a process that requires expensive machinery, leaving farmers at disadvantage and at mercy of middlemen charging for roasting services.
Solution: We propose the development of a low-cost, fully automated, AI-powered coffee roaster for small-scale coffee growers in Latin America. The roaster will allow farmers to produce different coffee specialties. For rural coffee growers, being able to roast their own coffee can mean an increase in value of up to 2 times compared to selling dried beans.
What is the problem you are solving?
Close to 125 million people worldwide rely on coffee for their livelihoods. Currently, coffee is the most important, widely traded agricultural product. In Colombia for example, more than 300,000 families associated with the National Coffee Federation, grow and sell coffee all year-round. However, the coffee market has been changing in the past decade giving pass to an increased interest in quality coffee. Producing quality coffee requires in many cases access to expensive infrastructure, as it is the case of coffee roasters. Given the fact that most of these small-scale farmers are not able to afford roasting technology, they are left to pay, more often than not, unfair fees for roasting services. Having to sell post-harvest dried beans means a farmer can only capture a third of the potential revenue they could gain from their crop. Having to pay for roasting services, means farmers cannot capture the entire value of their product due to technology inequity. It is precisely at this intersection where our solution sits.
Who are you serving?
Since the inception and main driving force of this project come from the farmers themselves, the short and honest answer is that they are serving themselves. What we do is collaborate with these small-scale coffee growers in rural Colombia whose main economic activities revolve around coffee. During the past year and a half, we have been experimenting together with ways in which we can address this challenge. Through co-design workshops with local students and professionals, as well as with international students (including MIT), we have advanced a first prototype. The process leading to this prototype has been focused in augmenting the already existing experimentation processes from farmers in the Cundinamarca region. This approach to co-design and co-production is allowing farmers to increase local capacity for technology design and manufacturing, as well as supporting a mindset of self-determination in their communities. Most importantly, their continuous involvement has made it possible fort the technology to be a direct response to their pain points.
What is your solution?
A traditional 10kg capacity coffee roaster costs around $5,000 USD in Colombia depending on its quality. This price is fairly homogeneous across Latin American countries. Our solution is a low-cost, AI-powered coffee roaster that costs around $1,000 and can be bought in monthly installments or built entirely by farmers themselves. The roaster has a capacity of 10kg per cycle, and will run for about 30 minutes per cycle. The roaster includes integrated cooling and extraction systems, a stainless steel drum, a thermocouple, a sampling chamber and a Raspberry Pi based automation and learning system. This system allows farmers to select among different roasting curves leading to different coffee varieties. Once a curve has been selected, the system controls the roaster mechanisms accordingly. The farmer can provide feedback to the roaster via a screen and the system utilizes this information to improve its performance over time. For farmers who are open to experimenting, the system can generate new roasting curves through a generative algorithm and produced new ones as it gathers feedback from farmers.
The entire design of the roaster will be open-source, with a paid modality for farmers and hobbyists who would like to purchase the roasters instead of building them.
Select only the most relevant.
Where is your solution team headquartered?Fusagasugá, Cundinamarca, Colombia
Our solution's stage of development:
Select one of the below:
New application of an existing technology
Describe what makes your solution innovative.
There are three key innovative aspects to our solution:
- By using open-source hardware and software, we are significantly lowering the cost of production of what constitutes a key technology to enable rural coffee-growers to increase revenue.
- We are inverting the direction of the relationship these farmers have with post-harvesting coffee technologies by allowing them to hold control over the design and production of the roaster.
- By using machine learning technology, we will be able to create new AI-generated coffee roasting profiles, bringing an entirely new conversation to the table around the role of machine intelligence in agriculture.
Describe the core technology that your solution utilizes.
There are at least three technologies at the core of the roaster:
- A combination of digital and artisanal manufacturing capabilities. The body of the roaster is industrially designed; the cooling system for example, uses traditional pots and local manufacturing.
- All the roaster's control systems are built over open-source hardware and software. From Arduino all the way to TensorFlow.
- The roaster will integrate machine learning algorithms to generate new coffee flavors at the roasting stage. By doing this, our technology brings an entirely new concept to the industry: machine-generated coffee flavoring.
Why do you expect your solution to address the problem?
For the past 3 years we have been doing continuous work with coffee-growing cooperatives in the region of Fusagasugá, Colombia. This work has included a number of experiments and iterations of the roaster, all of which have resulted in a current prototype being used by one of our main partners. Through this work, we've learned the importance of this technology for farmers to be able to make the most out of their crops. In fact, when we arrived for the first time in this community, we found farmers already experimenting with roasters. Our work has been built along theirs.
In summary, if we make roaster technology available to rural, small coffee growers, they will be able to sell their crops at better prices because the market rewards specialty coffee. This will have as a result a direct positive impact in farmers revenue.
Select the key characteristics of the population your solution serves.
In which countries do you currently operate?
In which countries will you be operating within the next year?
Select an option below:
How many people work on your solution team?
4 people full time
1 person part-time
For how many years have you been working on your solution?
A year and a half
Why are you and your team best-placed to deliver this solution?
First and foremost, the core of our team are farmers themselves. One of our team members, Franklin Espitia, who is a farmer himself, constitutes the core of our team. He has been tinkering with roasters for many years, and as a coffee grower, understands the needs and desires of his profession, and his community.
Along Franklin, we have a team that combines industrial design, electrical engineering, architecture and anthropology. A group of people who have been working on this challenge for more than the year and a half since Franklin brought the first roaster to life.
With what organizations are you currently partnering, if any? How are you working with them?
Locally, we are collaborating with the following organizations:
- De Finca & APRENAT. Coffee growing cooperatives in the Fusagasugá region in Colombia. They provide us with technical advice and access to testing mechanisms.
- Universidad de Cundinamarca. Regional university supporting us with technical advice, access to infrastructure and human resources.
Abroad, we a recurrently collaborating with two organizations within MIT:
- MIT Media Lab
- MIT D-Lab
These two partners are providing us with technical support and advice.