15+ years of product leadership at Google and Meta — now helping startups ship AI products that matter.
I spent over 15 years in product leadership at Google and Meta, most recently as a Group Product Manager overseeing AI-powered product lines used by hundreds of millions of people.
At that scale, you learn things you can't learn anywhere else: how to prioritize ruthlessly when every feature seems important, how to take AI from a demo to a product that works at scale, how to build the monitoring and quality systems that prevent AI from embarrassing your company, and how to make product decisions when the technology itself is a moving target.
The Bay Area is full of brilliant AI startups with world-class engineering teams — but many of them are missing the senior product operator who's done this before. They don't need a full-time CPO yet, but they need someone who can walk into the room and immediately see what's working, what's not, and what to do next. That's what I do.
I can't share everything, but here's what shaped how I think about AI products:
Shipping AI at massive scale. I led product for AI-powered features that touched hundreds of millions of users. I learned that the product decisions around AI — what to automate, what to leave to the user, how to handle uncertainty — matter more than the model itself.
AI quality and reliability. I built a 4-step monitoring framework for AI chatbots that reduced overpromises from 15% to under 2% of conversations. Most startups don't think about AI quality until it's a crisis. I help them build it in from day one.
Managing through ambiguity. AI product development is fundamentally different from traditional software PM. The capabilities change monthly, the competitive landscape shifts weekly, and the "right answer" is often unclear. I've spent years making high-stakes decisions in exactly this environment.
Technology is the easy part. The hard part is figuring out what to build, for whom, and why they'll pay for it. Most AI startups fail on product, not technology.
Speed matters more than perfection. In AI, the landscape changes so fast that a perfect plan executed slowly is worse than a good plan executed quickly. I bias toward shipping and learning.
Demos don't equal products. The gap between "look what our model can do" and "here's a product someone will pay $X/month for" is where most AI startups die. I help bridge that gap.
You need a moat, not just a model. If your entire product is a thin wrapper on an API, you're one OpenAI release away from irrelevance. I help startups find and build defensible product advantages.