ML model creator app
The goal of this project was to make machine learning accessible and practical for business use. The primary challenge was designing a system that enables users to create data models of real-world objects using nothing more than a standard mobile device.
Role:
Product Designer at CareAR®
Activities:
Research, Prototyping, Testing, Visual design
All designs, concepts, materials, and solutions presented in this case study are the exclusive property of CareAR® and are protected under applicable intellectual property and copyright laws. The content is displayed solely for informational and portfolio presentation purposes to demonstrate professional experience and capabilities. Any reproduction, distribution, or reuse of this material, in whole or in part, without the express written consent of CareAR, is strictly prohibited.
Research
I started with discovery - understanding the problem space, user needs, and competitive landscape. With access to potential users, I conducted a series of interviews to better understand their current pains, needs and preferable workflow. After 8 user interviews and 2 stakeholder/development workshops, I proposed an MVP feature list.
not familiar with Machine Learning (ML)
Most participants had heard of machine learning but weren't clear on how it worked in practice. Which gave me a strong feeling that proper onboarding for the app will be needed.
used a 3D modelling software at least once
Mixed responses in interviews pointed to the need for multiple input methods - a simple mode and an advanced mode for creating 3D bounding boxes.
of potential users happen to be technicians
My target audience mostly consisted of technicians who were eager to make their workflows more efficient and were looking forward to incorporate modern technologies into their daily tasks.
Solution
Scanning flow
During scanning user is asked to walk around an object while the app is taking series of pictures which are later used as the basis for data model creation and ML training. To help user understand when it is enough to film specific surface of the bounding box we've introduced dynamic color filling, which visually shows which area is already covered good enough.
Model training
You can start a data model training process by selecting scans you've just did. System will update you on a status of each data model and will notify when training is done.
