Design concept
I led the creation of user flows and designed the interface for the Minimum Viable Product (MVP) of the web-app dedicated to streamlining the claims process for damaged cars. Collaborating closely with the founder, tech lead, and two machine learning engineers, I meticulously crafted every detail and interaction within the app.
The data inputted by users in the initial version of the app not only informs claim evaluations but also serves as training data for the machine learning models. Consequently, precision and adherence to consistent guidelines were necessary for each claim. The accompanying screenshots showcase the use of silhouette assistance to aid users in capturing comprehensive images of the vehicle.
User flow
Claim report
Following the data input process, we generated a comprehensive claim report where all damages were meticulously annotated and marked. This manual annotation process, performed by mechanics, serves as foundational training data for machine learning models. These models are trained to discern between damages and non-damages, identify various types of damage, and accurately categorize them.
