About
AI at enterprise scale.
I am a product and engineering executive in the media, streaming, and advertising space: currently Hearst Autos, previously Warner Bros. Discovery / Motor Trend, Pandora, Starz, Disney, Hulu, Apple, and Microsoft. Along the way: Microsoft Gold Star (2x) and Disney Innovation Award.
My focus is rolling out AI at enterprise scale across three layers. AI for customers: capabilities embedded in products. AI for employees: governed model access, internal platforms, and workflows. AI as the engineering factory: agentic systems that plan, build, test, and review software with human oversight.
The engineering factory is where I spend most of my time, because it is where the leverage compounds. The interesting problem is no longer whether an agent can write code. It is alignment, verification, and approval: making every agent change arrive as a working, isolated, testable environment with evidence, not a plausible diff.
I have done the pre-AI version of this transformation at scale: at Motor Trend we took release cadence from 18.9 days to about 10 minutes and rebuilt the culture around trunk-based, evidence-driven delivery. The agentic version is the same playbook with a new workforce. The full history is on the experience page.
I post one insight from the field every day on X. The longer versions live on the blog. If you are wrestling with an AI roadmap, an internal platform, or an agentic delivery architecture, get in touch.
Common questions
What does "AI as the engineering factory" mean?
It means redesigning how software gets built, not just adding assistants. Agents plan, implement, test, and open pull requests inside controlled pipelines, while humans define constraints, review evidence, and improve the system. Every agent change arrives as a working, isolated, testable environment rather than a plausible diff.
How is that different from giving developers coding assistants?
Coding assistants make individuals faster. A factory makes the organization faster. Assistants are stage one of a maturity model that ends with agentic delivery systems: isolated preview environments per change, independent verification, human approval gates, and measurable throughput, quality, and cost.