When the same customer looks different everywhere
Bringing billing, CRM, access, and support together so teams can act on one reliable view.
Systems builder and operator
I help teams spend less time fighting their tools and more time doing useful work.
My work sits between business operations and technology: understanding the real problem, connecting the right systems, and keeping the result dependable.
Understand before building
Map the workflow, find the expensive failure, choose the simplest reliable change, then measure it.
Notes from the work
Clear explanations, screen recordings, and lessons from building and operating real systems.
Featured case studies
Reliable systems that connect operations, data, AI and software
Four problems I solve again and again. Employer work is shown as the method on made-up data; every figure says where it comes from.
01Customer identity reconciliationOne customer, one recordPayments, CRM and community access, joined on one canonical identity.Read the case study 02Customer lifecycle automationProving a member was really removedOnboarding, offboarding and customer success steps that prove themselves.Read the case study 03Lead intelligence and revenue operationsFaster lead response, cleaner lead dataInstant lead response and cleaner outreach lists for two clients.Read the case study 04AI agent engineeringAI agents that have to prove their workAgent orchestration, tooling and validation in my own operating system.Read the case study
How the work holds up
Built like production software, because it is
01
Check the destination
A green run is not proof. I confirm the record changed, the alert arrived, the access was removed.
02
Refuse to guess
When the evidence is unclear, the system holds the case and says why. A person decides.
03
Gate live writes
Anything that changes a live system goes through review and approval first, and is read back after.
04
Measure it
Before and after, with the method written down. If it was not measured, it was a guess.
More of the work
The foundation underneath
- DevOps and cloud engineeringCI/CD, security controls and cloud infrastructure for a large retailer, plus personal cloud builds.Previous work
- Technical experiments and open sourceFrameworks, prototypes and tools built in the open or for myself.Projects
- Engineering background and educationFrom chemical and industrial engineering to DevOps to AI operations.About
Latest writing
Give your team a workflow, not another AI login
Why giving people AI access is not enough, and how one familiar task can become a repeatable way of working.
A successful run is not proof that a workflow worked
Why green dashboards, completed runs, and clean logs still do not prove that a business workflow produced the intended result.
Where AI-assisted workflows need hard boundaries
A practical way to separate model judgment from the deterministic checks and human decisions that keep a workflow safe.
Have a system that keeps drifting?