MealMatch AI

Inspiration

Reflecting on our daily eating habits, we realised that wasting food is not only a waste of the food itself, but of hundreds of litres of water that gets wasted during irrigation and ounces of efforts that farmers and chefs go through everyday. A single hamburger can waste up to 660 gallons of water. Food waste is easy to minimise when there is an appropriate system designed for efficiency, empathy and to work towards a circular economy. The concept of Circular Economy is yet to be implemented in our city, Ahmedabad, and this may be one step towards achieving the same. We realised how rare proper waste segregation is, and that free wet waste may support organic farming by providing a cheaper alternative to chemical fertilizers. We also realised that animal shelters, in our local context- gaushalas (cow shelters), rarely receive food donations. This helped us come up with a comprehensive solution, suited for our city.

What it does

This platform is acts as an intermediate between shelters, animal shelters and farmer-related NGOs by categorising surplus food into three categories- edible for human beings, edible for animals and used for compost. Through this, we maximise resource-use by minimising wastage. This website has 3 dashboards- admin, volunteer/shelter and restaurant/donor. Communication channels have been established to bridge these gaps. A map feature as well as 'Claim', 'Approve Image', filtered search etc. features have been added as well.

How we built it

We built this by researching edible-criteria for humans, animals and compost materials and surplus food habits. On a technical basis, we utilized our own coding skills including python on Visual Studio Code and used the help of Codex, GitHub Co-pilot and Claude to help us resolve bugs and link APIs. The usage of Hugging Face API was done for recommendations and classifications.

Challenges we ran into

One major challenge that we overcame was the AI-associated risk. Since food accidents can be fatal, especially for animals, we resolved this using a three-pronged verification stage- 1. using the AI system 2. Using a human-in-the-loop 3. User input verification. Most APIs were not available in our region, without a paid account, this task is still a challenge for us.

Accomplishments that we're proud of

We are proud of the fact that our idea may do real good and have a positive impact on our own community if we manage to efficiently execute it. This is a prototype, however, our actual MVP may be able to successfully create the culture of circular economy especially during festivals and wedding seasons!

What we learned

We learnt that programming websites for social causes is a challenging task requiring tremendous hard work and it made us realise that we should be more grateful towards developers who do this on a daily basis. We learned that food management systems are rare, but may work out when efforts are made, and that communities must join hands to solve problems locally. On a technological point of view, we sharpened our programming skills and learnt real-world applications of programming languages.

What's next for MealMatch AI?

We will be adding WhatsApp message features to notify farmers only when compost reaches a substantial amount. We plan on completely upgrading our user-interface, make it more user-friendly but also more secure with logins. We will add a better database management system for admins to add multiple roles. We are developing predictive modelling to create heatmaps for food surplus, especially during festival or wedding seasons. We also wish to add predictions based on expiry, carbon footprints and so on. Mostly, we wish to have real restaurants and shelters signed up for our MealMatch AI!