Inspiration
Embedded finance and open banking infrastructure talk by Nicolas Benady at the start of the hackathon.
What it does
Our application predicts inventory needs to reduce production downtime, optimizes purchasing timing, and secures the best offers using rich empirical data—all within a single, end-to-end platform from forecasting to payment.
How we built it
We built a full-stack application using React, TypeScript, and HTML/CSS for the frontend, and Python with FastAPI for the backend. We leveraged machine learning models for inventory forecasting, and implemented a transformer-based model to rank suppliers. For the payment infrastructure, we integrated Swann and Google AP2.
Challenges we ran into
Framing the idea effectively to align with current market needs was a key challenge. From a technical perspective, selecting the most appropriate model for each use case required careful consideration. Finally, integrating and connecting the different payment solutions also proved to be complex.
Accomplishments that we're proud of
We are very proud of the model we were able to train in order to predict when the client will need certain pieces.
What we learned
By framing the idea clearly from the outset, we were able to fully realize the benefits of our approach during development, iterating quickly thanks to well-defined objectives.
What's next for MRO Pilot
After this hackathon, we want to join Station F for one month to refine the technology behind the POC we have already built, launch our first paid pilots, and keep iterating with real industrial customers. From there, the goal is to scale the number of clients while advancing through our product roadmap, including Deep Research to find and compare new suppliers.
Log in or sign up for Devpost to join the conversation.