Case Study: Lovi.Care improves QA, data labeling, and algorithm updates with Toloka

Lovi.Care launches QA in 50 minutes with Toloka

Lovi.Care, an AI skincare app, needed a way to thoroughly test its high-risk face-scanning feature and collect diverse, real-world data to train its algorithms. Their challenge was that automated and internal QA testing could not reliably simulate the unpredictable conditions and edge cases encountered by actual users on a wide variety of devices. To address this, they turned to the Toloka Platform for help.

Using Toloka, Lovi.Care implemented a solution where real-world contributors installed and tested the app, providing recorded sessions that uncovered critical bugs and collected valuable data. This approach allowed them to perform rapid QA, target specific data for labeling, and efficiently calibrate algorithms after major updates without needing to relabel entire datasets. The results for Lovi.Care included drastically reduced setup time, with projects launched in under an hour, and all test scenarios completed within a day, ensuring a more stable product for their large-scale launch.

Lovi.Care