A deliberately small practice. I'm the only operator — by design. That means the person you talk to is the person doing the work. On AWS since it was basically just S3, now focused on shipping production agentic AI on Amazon Bedrock.
A snapshot of recent engagements. Names withheld where contracts require it, but the architectures and outcomes are real and in production.
Studio teams answer cross-system operational questions (project status, storage costs, timelines) in minutes instead of days. One agent pulls context across separate asset, project, and finance systems.
A supervisor agent on Amazon Bedrock routes plain-language questions to four specialist agents covering driver, asset, customer, and route data, replacing four separate dashboards with a single conversational interface.
Built to learn the GenAI stack on AWS, kept because it works. A Claude Code agent in ECS/Fargate picks up cards, does the research, and writes back, turning customer questions into reviewable drafts in near real time. About $13/mo.
A fixed-price sprint that ends with a working proof of concept and a clear, evidence-backed recommendation you can take to leadership.
Most GenAI projects stall in the "is this a fit?" phase for months. The Sprint compresses that phase into 15 working days, ending with something running in your AWS account, a cost model, and a written recommendation. No retainer, no commitment to the next phase.
Hundreds of customer deployments under my belt. The classics still hold. A solid AWS foundation makes the GenAI layer above it boring in the best way.
A secure AWS foundation deployed via CloudFormation. Multi-account, identity, monitoring, and a reference architecture of your choice.
02I plug into your existing team where you're stretched, from architecture through to keeping it running. Senior AWS capability, at a fraction of what hiring that depth in-house costs.
03A point-in-time review across IAM, networking, logging, data, and edge. You get a prioritized fix list and a report, walked through with me.
I'd rather build the right thing in three weeks than the wrong thing in three quarters.
Engagements are short, scoped, and outcome-defined. I'll tell you when something isn't a fit for AWS, or for Bedrock, or for me, long before you've spent money you didn't need to. The goal is your team owning what we build, not depending on me to keep it running.
Long-form posts about what I'm building and what broke along the way. No hot takes, no listicles.
Send a paragraph about what you're trying to do. I'll usually get back to you within a few days.
