What it does
TruckAI allows users to swipe left and right on trucks according to their liking. When the user swipes left and right, a Clarifai ML model is trained using the user's "right swipes." As the user swipes more the Clarifai API gets a better idea of the user's truck needs and shows the users images that match that model by a higher probability. At the same time, any user can create a new truck and put it up for sale. When a user uploads a picture, a general pre-trained Clarifai model makes sure that the image is of a RealTruck and not an SUV, bike, or anything else. Whenever a buyer swipes right on a truck they are given the option to make a monetary offer. Truck sellers can then swipe left and right on all payment offers. When a buyer and seller are matched, they are given access to a private chat room and a quick and easy payment system.
How I built it
I used Firebase and Google APIs for the backend and authentication. I used Clarifai for image recognition, Stripe for payment, and Android Studio for UI//building the app.
Challenges I ran into
Training the clarifai API with user data was completely unchartered waters for me.
Accomplishments that I'm proud of
Learning how to use the clarifai API.
What I learned
I learned some nifty UI design tricks and how to use Stripe and Clarifai.
What's next for TruckAI
I'm probably gonna expand the market to cover cars, bikes, and such and do some customer discovery.