A partial record of work I can talk about publicly. Several significant engagements — production AI systems for value-based care, petabyte-scale data platforms, public-sector infrastructure — aren’t represented here because the details are sensitive. I’m happy to discuss those directly under appropriate agreement.
Moving 24 Oracle instances onto a 13-site multi-master MariaDB topology that had to survive county-level network outages and keep elections running. Delivered three months ahead of a $12M contract. Still has 100% uptime through hundreds of election cycles.
Authoring the internal migration runbook for Amazon's company-wide mandate to move applications off VMs and onto AWS, in the era before CloudFormation, Lambda, ECS, or most managed services existed. The primitives were EC2, S3, EBS, and early RDS — and the architectural judgment we developed then is exactly what later abstractions let engineers stop thinking about.
Under appropriate agreement: Production AI patient outreach in regulated healthcare, clinical document intelligence using vector embeddings at scale, near-real-time change data capture for Snowflake, Fortune 100 database architecture for MariaDB enterprise accounts, and additional public-sector engagements. Get in touch if any of these are relevant to work you’re doing.