background
The agentic work is recent. What makes it deployable is not — it is two decades of migrations, compliance reviews and clusters that had to stay up. The governance instinct came from the boring half.
Cloud / AI Platform Architect · regulated federal program (client withheld)
Designed the governed agent execution architecture for a regulated federal environment. Governance as substrate rather than wrapper: on-behalf-of identity, tiered authorization, append-only hash-chained audit.
Founder · own practice
Independent platform and AI architecture work, and the home for the open-source tooling — OpenAgentics, Brainbow, Synth and PatchBay.
Principal Cloud AI Architect · research compute
Ran the research compute estate — a 256-node GPU cluster on Slurm and Kubernetes, with the MLOps tooling around it.
Cloud & infrastructure architecture · healthcare, enterprise SaaS, ISV
Enterprise Kubernetes adoption under HIPAA. OCI contract-to-cloud automation and customer onboarding pipelines. Two full datacentre-to-AWS migrations. The unglamorous half of the résumé, and the reason the new half works.
Five systems, five different reasons they exist — and the design decisions underneath.
See the work →