Engagements and delivered systems

GitLab CI pipelines for a HITRUST-regulated healthcare system, where every stage has to produce evidence as well as an artefact.

In a regulated environment a pipeline is not just automation. It is the control that proves how software reached production.

Training, registry and deployment workflows on SageMaker for a healthcare-compliant environment, where a model version has to be traceable to its data.

A model in a regulated setting needs the same provenance as a release: which data, which code, which parameters, approved by whom.

Containerised ML training and inference workflows served from K3s, with monitoring that treats a model service as an ordinary production service.

Model services fail in ways that HTTP 200 does not capture. The platform had to make that visible.

Terraform-described, cost-conscious AWS architectures with least-privilege IAM and segmented networks, built during the Adex apprenticeship.

Availability, cost and blast radius are the same conversation. Designing one without the others produces an architecture that only works on a slide.