All success stories

AI App Building & Hosting · Success story

Taking an AI app-building platform from on-premises to AWS

CloudVests migrated BuildX AI's platform to AWS, running its AI build agent on Amazon Bedrock and replacing third-party sandboxes with a Spot-backed EC2 fleet.

logoClientBuildX AI
0third-party sandbox dependencies
100%environment deployed with AWS CDK
Multi-AZencrypted database and cache
30%cost per AI build, or cutover downtime

Selected AWS services

Core services in the architecture.

A focused view of the AWS services most representative of this engagement. The complete technical scope is described below.

  1. 01Amazon ECS
  2. 02Amazon Aurora
  3. 03Amazon ElastiCache
  4. 04Amazon EC2
  5. 05Amazon Bedrock
  6. 06Amazon Bedrock AgentCore
  7. 07Amazon SageMaker
  8. 08Amazon CloudFront
  9. 09AWS CDK
01

The challenge

BuildX AI's app-builder platform ran on-premises and relied on a third-party sandbox service for AI code execution, while its margins depend on the infrastructure cost of every build.

02

What CloudVests delivered

CloudVests migrated the platform to AWS on an AWS Control Tower landing zone, with the entire environment defined in AWS CDK and released through AWS CodePipeline.

Java and Node.js services run on Amazon ECS with AWS Fargate. PostgreSQL moved to Amazon Aurora and Redis to Amazon ElastiCache for Valkey, both Multi-AZ and encrypted. A self-managed fleet of isolated EC2 sandboxes, mixing Spot and On-Demand capacity, replaced the third-party service.

The AI build agent runs on Amazon Bedrock and Amazon Bedrock AgentCore, with Amazon SageMaker hosting self-managed models. The React frontend moved to Amazon S3 and CloudFront, keeping custom domains and preview links.

03

The outcome

BuildX AI runs its full platform on AWS with no third-party sandbox dependency, and Spot capacity keeps the cost of each AI build down.

Multi-AZ data, encryption and compliance evidence give it the foundation to sell to enterprise customers.