Key results
The company
isento
isento.deIT services provider for software development, quality assurance, and business analysis.
Result highlights
- 85% reduction in storage costs
- Under 10s to clone 30,000 images for users
- 85% less storage space vs traditional duplication
The story
A German IT services provider specializing in tailored software engineering and complex project delivery for industrial innovation.
Clients needed to rapidly iterate on visual AI models, but traditional workflows required duplicating massive image datasets for each experiment. This process generated prohibitive storage costs and slow development cycles that prevented non-experts from adopting the technology.
The provider developed a no-code platform using Vertex AI for model training and Cloud Run for serverless scaling. BigQuery handles the data layer by utilizing zero-copy cloning to replicate dataset metadata instantly without duplicating the underlying files. Artifact Registry stores the final containerized models, enabling a workflow that scales automatically from single users to thousands.
Scope & timeline
- 100+ schools reached with AI robotics platform
Quotes
“Vertex AI provides an integrated training environment for seamless model creation and deployment, while BigQuery gives users the freedom to use large sets of data. That means our customers can immediately start using the platform and solving problems with AI.”