I start with a business problem that happens to need a platform underneath it, which means locating the moat, naming what caps it, and choosing the adjacency the business has earned the right to take next. Architecture is a capital allocation decision written in code, and it outlives the strategy that funded it. Over two decades I have built the platforms large products run on and the organizations that run them: teams of 200 and up across multiple countries, P&L ownership above $20M, acquisitions absorbed, and cloud, data and AI transformations at Shutterfly, Apollo Education, Pearson and Wood Mackenzie. I set the direction and the architecture, hire and grow the leaders who own delivery, and go deep myself on the decisions that are expensive to reverse.
Bylined on Amazon's own engineering blogs, and the pattern behind them now built into Microsoft Copilot.
Work referenced in company and industry coverage. Click any item for how it maps back to the work above.
The agentic platform and MCP-based agent integration behind Synoptic, covered in more technical depth in the Machine Learning Blog post above.
Read the sourceThe student and instructor AI tools described in the four-platforms blog post, now shipped across Pearson's courseware.
Read the sourceThe Cognito-based identity platform and migration I led, detailed in the identity-at-scale blog post.
Read the sourceAn early look at the adaptive learning platform work that began during my time at Apollo Education, University of Phoenix's parent company.
Read the sourceA vertical search feature I built early in my career, embedded directly into the trip-planning workflow rather than sending users elsewhere.
Read the sourceIPTV middleware and scale-benchmark work from earlier in my career, before the move into education and enterprise platforms.
Read the sourceEight strengths I bring at scale, and underneath them the tools and technologies I have actually run in production.
Where I go deep, what I hand over, and what I stay accountable for either way. I still operate at every level of this stack.
Filterable by skill, the way a language model loads a skill file. Pick a lens and only the work that matches stays on the page.
Wood Mackenzie2025 to nowVP, Architecture, Platform Engineering and AI
P
Pearson2020 to 2025VP, Architecture, Platform Engineering and AI
S
Shutterfly2015 to 2020Senior Principal Architect
AE
Apollo Education2009 to 2015Principal Solutions Architect
IO
IoTangoFounderCo-founder, AI-connected IoT
+3
Travelmuse, Kasenna, Hewlett-Packard2002 to 2009Engineering, IPTV middleware, J2EE
A layered, distributed reference architecture built AI-first. Applications and an API layer sit over a model-agnostic agentic framework, hub-and-spoke MCP, advanced RAG and MLOps, all on governed data foundations and shared platform-engineering infrastructure. Twelve interactive architecture diagrams, spanning agentic AI, data, distributed systems and modernization, and my platform first principles live on the dedicated page.
Across AI-driven learning systems and distributed systems. Most came from building adaptive education platforms, plus a 2025 patent on AI-generated adaptive course templates. Each one is explained in plain language on the dedicated page.
Standing up agentic AI, modernizing a data platform, or setting the governance that makes it responsible. I am always glad to compare notes with other builders and leaders.
Built with AI, under my review. I build AI platforms, so I used them here. LLMs wrote much of the markup and styling; the architecture, narrative, experience, and every word are mine. The assistant runs in your browser: nothing sent, nothing logged.