AI Pipelines · Design Engineering · Creative Operations
Machines handle the routine.
Humans build what matters.
We design and deliver systems that remove operational drag so people can spend their time on judgment, taste, and original work.
What We See
Does this look familiar?
01
Creative teams burn hours on work that a pipeline handles in seconds.
02
The automation breaks at 2 AM. Nobody knows why. No fallback exists.
03
Fourteen AI tools, zero coherent system connecting them.
04
One team adopted AI for image gen. Everything else still runs on copy-paste and Slack messages.
05
Tribal knowledge is the workflow engine. When someone leaves, the process leaves with them.
Services
What I do.
AI Pipeline Design & Build
Best for: Teams with manual workflows that AI can replace end-to-end
Deliverables
- Working pipeline
- Integration map
- Fallback paths
- Monitoring setup
- Runbook
You get a pipeline that runs, not a deck that describes one.
3–6 weeks
Operations Automation
Best for: Ops teams still running on handoffs, spreadsheets, and Slack
Deliverables
- Workflow audit
- Automated handoffs
- Human override paths
- Rollback patterns
The process survives someone going on vacation.
2–4 weeks
Enterprise AI Governance
Best for: Regulated orgs shipping AI into production for the first time
Deliverables
- Risk map
- Evaluation framework
- Audit trail design
- Policy controls
Your AI can explain itself when compliance asks.
2–4 weeks
Open Source
Tools I build and share.
Coming soon
We love open source!
How I Work
Ship small. Measure everything.
Diagnose
I sit with the team, watch the actual workflow, and find where the hours go. No questionnaires.
Design
Architecture and control surfaces. What the system does, where humans stay in the loop, how it fails safely.
Build
Working code in your environment. Not a prototype on my laptop.
Hand off
Your team runs it without me. Docs, monitoring, the whole thing. I leave and it keeps working.
Proof
Where this work comes from.
Console Development Platform
Oracle Cloud — UX infrastructure enabling 100+ teams and 150+ JS products
Platform-scale design consistency and delivery velocity
Identity & Access Management
AWS IAM — policy tooling and identity UX
Usable security at enterprise scale
Cloud Financial Governance
AWS Billing / Cost Management — Insights, Budgets, Organizations, Apptio context
Measurable cloud cost governance
Enterprise Collaboration Suite
AWS WorkMail / WorkDocs — admin and end-user UX
Coherent enterprise collaboration tooling
Voice & Calendar Interaction
Alexa for Business — group, time, and calendar management
Enterprise voice interface patterns
Infrastructure Engineering
Istari Digital — infrastructure deployments, Lockheed Martin Skunk Works partnership
Production infrastructure at defense-grade standards
Point of View
Working tenets
Designing AI Systems That Survive Audit
Most AI systems can't answer the question "why did it do that?" Here's how to fix it before the auditor asks.
ReadHuman Override Patterns for Agentic Workflows
What happens when the agent does something wrong? If you don't have an answer, read this.
ReadEvaluation UX: Traces, Scorecards, and Policy Controls
Traces, scorecards, policy controls. The UX that sits between your model and your org's trust.
ReadAtomic Design for AI Brands
Brand systems for AI products. Tokens, components, the whole stack.
ReadLocal LLM Infrastructure for the Solo Operator
Running your own models on your own hardware. What it actually costs, what works, what doesn't.
ReadThe Ratchet: Why Skills Beat Prompts
Prompts are disposable. Skills compound. The difference matters more than most teams realize.
ReadAbout
I have spent twenty years working on everything from streetwear sites to consumer software to enterprise platforms to software deployed to top secret environments. Some highlights include Oracle Cloud console infrastructure serving 100+ teams, cost management and IAM at AWS, Alexa for Business, and R&D at British Telecom. Three US patents came out of that work. I still have time to illustrate for skateboards.
The last three years, I have been exclusively focused on building AI pipelines and automation for creative organizations and maintaining a few open source tools I am releasing here spring 2026. I take a handful of engagements at a time and do the work myself.
Currently based in Seattle Area.
Patents
US 10,430,504 · US 9,967,332 · US 12,259,995
Got a workflow that needs fixing?
Thirty minutes. We look at what you have, I tell you what I'd build. No pitch deck.
Book a call