Case Study: Standard Cognition achieves rapid prototyping and real-time checkout accuracy metrics with Rockset

Standard Cognition Uses Rockset to Deliver Data APIs and Real-Time Metrics for Vision AI

Standard Cognition, maker of an AI-powered autonomous checkout that uses overhead cameras to track shoppers and purchases, faced a data challenge: massive, heterogeneous streams (video-derived events, transactions, inventories, vendor feeds) in different formats and schemas that made ETL, real-time accuracy checks, and fast prototyping slow and costly. To overcome this, Standard turned to Rockset as a fully managed, schema-on-read analytics platform for prototyping and internal analytics.

Using Rockset’s schemaless ingest, Write API, SQL engine and client libraries, Standard streams diverse data sets into a single queryable system, runs ad hoc joins and real-time accuracy metrics, and iterates features quickly—converting mature prototypes into serverless production endpoints. Rockset enabled Standard Cognition to productionize several endpoints a day, give over two thirds of its technical team fast access to data, reduce infrastructure overhead for experiments, and accelerate improvements to checkout accuracy and product development.

Standard Cognition

Tushar Dadlani

Computer Vision Engineering Manager