When we first released Fritz AI Studio earlier this year, we firmly believed in its technical backbone. Using the Studio, you could synthetically generate labeled datasets, use those to train mobile-ready models without code, deploy them across mobile platforms, and set up pipelines to collect real-world model data, making it easier to improve models over time.

And with a lot of help and feedback from the many conversations with members of our Fritz AI user community, we’ve added the connective tissue to that technical backbone that allows users to more seamlessly build and deploy models.

For our first big UX overhaul, we focused on the following:

• Account and project creation flows
• Project dashboards
• Clarity around model building steps
• Image annotation workflow

I’ll also highlight a couple of smaller improvements that smooth the edges of the model training job workflow.

To show these improvements in action, I’ll work through a simple demo project: building a custom golf ball object detection model.

#fritz-product #ux #heartbeat #mobile-app-development #machine-learning