The same problem in five different costumes.

I started in data warehousing in 2000, which means I have spent my whole career on one problem wearing different clothes.

The first version was getting data into one place at all — a senior developer role at Silicon Laboratories in Austin, then data warehouse content development at SAP Labs, then a data warehouse and data quality program at Boeing. Boeing taught me the thing that has held up ever since: a technically correct model nobody trusts is worth nothing.

The second version was making that data legible. At Accenture I led a cross-functional team of eleven turning a client’s legacy reporting into SAP Analytics across logistics, procurement, and finance. The piece of work from that period I still bring up is a distributor-spend solution that freed up to 20% of the client’s distributor capital, which went back into sales promotion. It was not a sophisticated model. It answered a question nobody had previously been able to ask.

Then HANA arrived and speed changed which questions got asked. At IBM I owned analytics architecture and the commercial side of delivery — cost structures, deliverable tracking, risk, RFP responses — and won and delivered implementations at Honda Motors, VSP, and Abbott. After that came cloud, and migration stopped being a project and became the transformation itself. Back at SAP I led analytics delivery for utilities clients under MaxAttention through their S/4 journey, drove mass on-premise-to-cloud migrations on HANA, and established a Cloud Database and Integration practice that still contributes recurring revenue.

Since 2022 I have owned an $18M portfolio of enterprise SAP cloud accounts end to end — deployment, adoption, renewals, expansion. Which brings me to the fifth version, and the one I find genuinely thrilling.

I implemented SAP Joule and guided customers through rolling out Joule Agents. A real share of my week now goes to questions that did not exist three years ago: how agents orchestrate, what they are permitted to ground on, where a human stays in the loop, what the blast radius is when one of them is confidently wrong. I delivered a Snowflake Cortex engagement connecting SAP data to LLM-powered analytics. I ran the workshops that moved PI onto BTP Integration Suite and BW onto Datasphere — the unglamorous prerequisite for all of it.

The pattern I have now watched five times is the same. The capability shows up, everyone is impressed, and then it sits there. What moves it is someone who can hold both ends at once — the architecture and the business case — and who is still in the room at the QBR when it is time to prove the number. That is the job I have. After eighteen years it is the most interesting version of it I have had.