SUPA partnered with a mobile app design company (the Client) focused on curating AI models to generate mobile app interface designs for their customers.
The Problem
The Client needed to expand the repertoire of their mobile app screen and interface design database for their AI models to train on. In doing so, the Client faced challenges in scaling up due to the many unique interface categories - a total of 117 of them - that the database contains. Given this fact, along with the high number of interface images to be classified, the internal annotating team the client employs encountered capacity constraints in conducting the classification work.
The Solution
The Client partnered with SUPA and utilised our highly-trained and experienced labelling workforce to ease the load on their internal team. SUPA's proprietary platform enabled a workflow that streamlined data classification tasks and quality checks, ensuring consistent data throughput while upholding the Client’s accuracy standards.
The Result
SUPA has allowed the Client to scale up their classified interface design database by 16 times within a 3-month turnaround time at 90% accuracy. Not only has SUPA met the Client’s agreed service accuracy level, SUPA allowed the Client to relieve internal capacities and reallocate resources to refining their AI models instead.
.png)
.png){kind=small}
.png){kind=small}