Case study
Analytics Workload Migration from AWS to GCP
How SquareShift moved a fast-growing EdTech platform to a fully serverless Google Cloud analytics stack.
A fast-growing EdTech platform, generating learner data faster than its stack could turn it into insight.
It needed a near-real-time, large-scale data platform to consolidate and enrich that data for analytics and ML, at a lower cost than the AWS stack it had outgrown.
Lower total cost of ownership. Data freshness, down from hours. Serverless Google Cloud stack.
Key servicesHow SquareShift delivered it.
The challenge
The platform ran its analytics on an AWS stack that had stopped keeping pace. Pipelines were batch-only, so business data was already hours old by the time teams could query it.
Machine learning and advanced analytics had no real home, and total cost of ownership had climbed above target. The client needed to consolidate and enrich data in near real time — for less.
What we delivered
SquareShift rebuilt the analytics workload on Google Cloud, fully serverless. Real-time ingestion runs on Kafka, with enrichment in Spark, replacing the batch loads end to end.
A governed, centralized data lake now consolidates and enriches business data for both analytics and ML. A detailed TCO analysis sized the platform and justified every part of the move.
The payoff
Data freshness dropped from hours to under five minutes, so decisions run on current data. The serverless stack cut total cost of ownership by 40% against the prior AWS setup.
Machine learning moved into production on the new platform, and the client now scales analytics without scaling the operations burden behind it.
Anyone can lift a workload into a new cloud. The real work is making it serverless and production-ready for ML — that's the difference between a migration and a modernization.
Analytics Modernization Practice Lead, SquareShift
Where this work sits