ProbiusDx
A legacy bioanalytical platform moved from Azure to AWS with 35% less operational overhead
The Problem
ProbiusDx ran a bioanalytical platform on Azure. It had grown expensive to operate and hard to extend. The business wanted it on AWS, with a stronger security posture and faster ML inference.
Constraints
- Labs depended on the platform every day, so the move could not interrupt them
- Compute-intensive analysis workflows had to keep running
- Security had to improve during the move, not after it
What I Did
- Designed REST APIs and microservices that wrapped the legacy components, so they could move one at a time
- Rebuilt the infrastructure on AWS with Docker and infrastructure as code
- Chose instance types for the analysis workloads and added auto-scaling based on real usage
- Hardened authentication, secrets handling and network boundaries during the move
- Ran old and new side by side until each service was verified, then cut over
Result
35%
Less Operational Overhead
Lower
ML Inference Latency
0
Lab Interruptions
The platform runs on AWS with 35% less operational overhead, a stronger security posture and lower inference latency. The labs using it were not interrupted.
Stack
Python
FastAPI
AWS
Docker
Kubernetes
Terraform
PostgreSQL
Jenkins
Planning a Move Between Clouds?
Tell me what runs where today. See how I work on systems integration and cloud infrastructure.