Loading
Senior Manager, Data & AI Products and Analytics Engineering at Disney Studios. I lead a matrixed data and AI product engineering organization — product managers, analytics engineers, technical writers, and program management — that runs enterprise-wide across the studios, and set direction for a wider footprint through governance forums. I brief the CTO and President of Studio Technology and Operations monthly, and co-lead enterprise AI task forces across product, program, and data.
Most product leaders manage. I run AI-native product teams and I ship real software. The combination is the point.
In parallel, a one-person software practice spanning household software, geopolitical intelligence, multi-household trip planning, multi-agent editorial infrastructure, forecasting and calibration, spec-as-code tooling, and AI prototyping. Every one carries a literal status, and some of those statuses are unflattering on purpose.
AI is what makes the combination possible. I am not working twice as hard — I am working differently, and the operating model is the same in both places.
The same operating model runs a full product organization at Disney and a one-person practice at home. Same primitives, different scale.
Roles, not prompts
A principal PM and a principal PMM review every spec before it becomes code. They hold standards, not context — which is why the output stays consistent across sessions that share no memory.
Specs under version control
PRDs, decision logs, and conventions live in the repo with the code they govern, validated on commit. When a decision gets revisited, the reasoning is still there.
Shipping as the forcing function
Typechecks, linting, test suites and content smoke tests gate every deploy; cost guards and liveness checks watch the data collectors. The discipline that makes enterprise platforms trustworthy is the same discipline, just without an org to enforce it.
Every figure carries its measurement method. If it can't be sourced, it isn't here.
~4 → ~1
Months from spec to shipped
Disney Studios · spec-to-ship duration, before and after the operating-model rollout
100%
Weekly active PM adoption of AI coding tools
Telemetry-tracked, not self-reported
80%
YoY loyalty growth on the Yum CDP
Marketing-mix-model attributed
3
Enterprise data platforms built 0 → 1
Financial services · restaurants · entertainment
Investment analysis into business intelligence into data products into AI product leadership. The domain kept changing; the work of turning messy data into something people trust did not.
Shipping is the forcing function. Understanding is the compounding asset. The products exist because building them is how the thinking gets tested.
Most product debt is decision debt. Make the decision, write down why, move. The decision log is worth more than the roadmap.
Technical decisions are organizational decisions in disguise. An architecture nobody can operate is not an architecture.
Open to Director and VP roles in AI product leadership.
ethan.c.stuart@gmail.com