About

I write about building reliable systems—lessons from two decades in product and operations roles where reliability mattered more than novelty.

Experience

Amazon — Built and scaled Amazon's product-safety systems used across global retail. Delivered the ML models and operating architecture behind 40M+ weekly interactions in 54 languages. Presented predictive safety controls to the Board and named inventor on U.S. Patent 10,223,353 (Amazon's Choice).

StockX — Delivered the company's first multi-year fulfillment roadmap and launched Express Ship—a profitable business line that cut delivery times from 9 to 3 days and reduced support costs by 75%. Rebuilt the product and engineering operating model to accelerate experimentation and reliability.

Stripe — Led global onboarding compliance and merchant-risk systems—handling hundreds of billions in payments for millions of businesses. Built an LLM-powered policy-evaluation engine that replaced hundreds of brittle keyword rules, reducing manual reviews by 80% and delivering Stripe's multi-year compliance automation roadmap.

I've spent much of my career designing systems that make organizations more reliable—whether for global marketplaces, supply chains, or financial infrastructure.

Now

I'm working on Flocksynthetics, building reliable AI systems for consequential work. The focus is the hard part between a compelling demo and a system people can trust in practice.

I write about what I learn: the tradeoffs, the failures, and the patterns that actually work.

I also advise founders and leadership teams on AI strategy, product direction, operating models, and reliability. That work ranges from helping solo founders turn an early idea into something real to helping companies with billions in annual revenue make AI useful at scale.

Advising

The work is usually some mix of product judgment, operating rigor, and translating what AI demos well into what actually holds up in practice.

I work with organizations at very different stages—from solo founders finding the right problem to solve to leadership teams bringing AI into large, complex companies.

That can mean sharpening a product strategy, pressure-testing an operating model, helping a team move from prototype to production, or identifying where AI will create real value—and where it will not.

If something on the site overlaps with what you're working on, LinkedIn is the easiest way to reach me.

Topics

AI systems: Practical reliability. How to build AI that leaders can trust.
Leadership frameworks: Decision-making systems that scale. Lessons from leading teams through hypergrowth.
Software craftsmanship: Architecture that lasts. The unglamorous work that makes systems dependable.

Why This Site Exists

Most writing about AI and leadership is either academic or promotional. This site is neither.

I write about what I'm actually building and learning—no pitch decks, no hype, just clear thinking about hard problems.