← All capabilities
Deep Learning
Purpose-built networks cover fraud, churn, and recommendations when tabular ML is not enough. They run as API jobs the same way the rest of the platform does.
Fraud and churn endpoints encapsulate training and scoring patterns teams otherwise rebuild. Neural-network and collaborative-filtering routes cover richer interaction data.
Use them from the dashboard or as a pipeline/agent step after ETL has landed a clean source.
What this does for the business
- Score transactions or accounts without a dedicated ML platform team.
- Feed churn risk into the same reports the supervisor already writes.
- Keep training bounded to the API’s job model rather than a hidden notebook.
API
POST /api/models/neural-network POST /api/models/fraud-detection Full reference: api.dsnresearch.com/docs