WildSnacks

Pushing the frontier of the SMB vending machine industry.

Comment

Inspiration

One of our team members runs a small vending machine business here on campus. Rather than being able to focus on growing his business, 90% of his operating time is eaten up by long and inefficient trips to and from the vending machine to restock it.

What it does

  1. Predict the price in relation to the local market and maximize profit and sustainability
  2. Predict when the machines or warehouse needs to be restocked based on warehouse and machine data
  3. Provide a route planning tool for SMB vending machine owners and operators/employees to refill vending machines economically and efficiently.

How we built it

We started by creating a single landing page to create a unified UI, and branched off, working on different components of our tool. These components were the login page/auth, the Warehouse tab, and the Route Planning tab. The three of us all worked on the full stack, developing frontend components, backend logic, and solving many server-side bugs.

Challenges we ran into

We struggled a lot when it came to dealing with bugs that appeared in deployment but not in local instances, as the files we used for API calls differed. Rate limits were also a challenge and slowed us down. Additionally, it was challenging to find a database server that fit our needs, especially because our tool's functionality heavily relied on many different spreadsheets of machine and warehouse-related data. The person who was in charge of designing the warehouse tab had minimal experience with backend and spent the majority of their coding time learning how to correctly simulate data, set up the backend, and connect it to the frontend.

Accomplishments that we're proud of

  • Being able to work with friends
  • Successfully debugging Authentication process (3 hours)
  • Successfully debugging backend integration with frontend (6 hours)
  • Not giving up
  • lots of AI!

What we learned

  • How to rapidly iterate on a product, going from MVP to a robust dashboard
  • How to use various tools, ranging from Vercel deployment, to authorization tools like Clerk, and even databases like Supabase.

What's next for WildSnacks

  • Restocking calculations at the item level (calculating which vending machines to restock based on when specific items are likely to sell out)
  • Determining which items to bring to restock vending machines
  • Navigation tools for drivers
  • Creating an interface for drivers, as ours is admin-centric

Vercel (hosting the website) Clerk (user 2FA authentication) React js FigJam AWS Cloud Three.js Supabase (DB) tesseract.js (OCR) (terrible so discarded) Eleven Labs (cookie AI)

Share this project:

Updates