← Case StudiesProject Summary
04
HealthcareCloud and Platform Engineering

HIPAA-compliant AWS infrastructure, rebuilt from a standing start.

Full infrastructure assessment, Terraform rebuild, and runbook delivery for a healthcare IT team ahead of a compliance audit.

Stack

AWS (VPC, EC2, RDS, S3, IAM, KMS)AWS CloudTrailAWS GuardDutyAWS Security HubAWS ConfigAmazon CloudWatchTerraformAWS Systems Manager

The Problem

A healthcare IT team had accumulated an AWS environment over several years that had never been audited for HIPAA compliance. PHI was stored in unencrypted S3 buckets. CloudTrail was disabled in three of four regions. IAM roles carried AdministratorAccess assigned to application workloads. RDS instances were unencrypted and running in the default VPC. There were no runbooks for the on-call team. An upcoming compliance audit had created urgency, but the team did not know the full scope of remediation required.

The Result

Passed the HIPAA compliance audit conducted three months after handover. Client IT team operates the infrastructure independently using the Terraform state and documentation delivered at project close.

What we built

The work that shipped.

Full infrastructure assessment documenting every compliance gap with remediation priority and owner
Rebuilt environment in Terraform: encryption at rest and in transit, CloudTrail enabled across all regions, VPC segmentation, and least-privilege IAM
CloudWatch dashboards and alerting configured for the on-call team before handover
Runbooks covering the ten most common incident scenarios, reviewed with the client engineers

Approach

How we designed the system.

The project began with a three-week assessment phase that produced a complete inventory of every compliance gap with remediation priority and owner. Thirty-one distinct findings were documented, ranging from critical (PHI in unencrypted S3) to informational (missing resource tagging). The rebuild was executed in a parallel environment — a new VPC and account structure was built in Terraform while the production environment remained operational. Workloads were migrated service by service, with each migration including a brief maintenance window for RDS cutover. The most complex phase was IAM remediation: 40+ roles were inventoried, mapped to their consuming workloads, and replaced with least-privilege service roles. The final environment is defined entirely in Terraform, with state stored in S3 with DynamoDB locking and a dedicated IAM role for pipeline execution.

Full technical report

Want the full
technical detail?

The full write-up includes architecture diagrams, technology selection rationale, implementation phases, and lessons learned. Available for qualified enquiries.

Architecture diagrams for the full system
Technology selection rationale and trade-offs
Implementation phases and key outputs
Lessons learned and what we would do differently

Available under NDA. We respond within two business days.