Four areas, one operator.
Technical leadership, AI systems, operating systems for teams, and the whole estate underneath. Most engagements touch more than one, and all of them start the same way.
Technical leadership
Architecture, hiring, unit cost
You have engineers and nobody setting the standard. I take the architecture and delivery decisions, so the pieces fit and the same problem stops being solved four different ways. I hire against an assessment I wrote, and I train the engineers you already have.
Architecture and delivery standards across every system the team ships
Technical assessments that measure how someone directs AI inside an existing architecture
Unit cost ownership, with a ceiling and an alert before anything switches on
Training for your engineers, not a dependency on mine
AI systems
n8n, Python, Claude
Decisions live in code and judgment lives in prompts, never the reverse. Every build ships with a test suite, a monthly running cost you see before it starts, and a handover pack your team can run without me.
Document and PDF extraction with multi-pass consensus and per-cell source verification
Agents wired into your CRM, inbox and ops stack, with approval gates on expensive steps
Churn recovery, inbound triage and lifecycle flows triggered on real signals
Content pipelines priced per finished post, with the monthly bill shown before the build
Operating systems for teams
Notion, ClickUp, CRM
One source of truth, and capacity you see before it breaks. Sized to your team rather than a template, and built so your operators own it after handover.
Custom Notion or ClickUp workspaces with role-based views
Roadmap and scope-discipline structures product teams stick to
Reports where every number links back to the row it came from
Documentation, runbooks and onboarding paths your team owns
The whole estate
Access, security, and cost per unit
The parts nobody owns until they break. Account migrations, credential rotation, offboarding, security posture, and what every running system costs per unit. This is where a diagnostic usually finds the money.
As-is architecture maps of systems nobody has documented
Account ownership, credential rotation and offboarding that survives a departure
Return-per-workflow audits, with the negative ones decommissioned
Self-hosted infrastructure on your own servers, with encrypted off-box backups
It starts the same way every time.
Read-only access for two or three days, then a written recommendation. From there we decide whether there is a build, a partnership, or nothing.
Fixed price on a build, or a monthly partnership. Never an hourly rate.
Have a problem in one of these areas?
Tell me what is going on and I will tell you what I would do about it.