Meet Our Founder

The Machine (Code and Theory). I was the technical owner of The Machine, the adTech industry's first Agentic Operating System, unveiled at CES 2026. I created the architecture, hired the team, and enabled the functionality, then led the 40-person global engineering organization that built it. The platform was fully built and deployed on GCP, using LangGraph and LangChain for agent orchestration with multi-tenant vector tables behind the RAG pipelines. I also developed the IP behind our most advanced features, including episodic memory and an adversarial reflection loop, which I prototyped first with Cursor and Claude Code so the team could build from working reference code. Governance was designed in from the start, with RAGAS for evaluation and Cloud DLP for PII. I worked directly with C-level executives and was consulted on mergers and acquisitions. I also created a Forward Engineering practice so client implementations of RAG and agents could be customized without forking the platform.

Interpreted Reality today. I'm a full-time Agentic Architect at Interpreted Reality, the AI engineering practice I founded. My work centers on agents and agentic loops, and on AI transparency: provenance for requirements, traceability across tests and releases, and governance sign-offs that can be audited end to end. Flywheel Forge is an internal product for creating and managing Interpreted Reality's products. In it, discrete agents pick up work, design, write, and QA code while reading from and writing to a Linear kanban board. Connect the Dots is a GraphRAG system for investigative reporting. It uses Claude Code with chained context windows and sliding retrieval to surface non-obvious entity relationships across large unstructured document sets. Both run on a custom Skills, agents, and MCP server framework that I designed to extend Claude Code for multi-step agentic workflows. I also advise enterprise firms on GenAI strategy, LLM governance, and production readiness.

Why it holds together. The through-line across twenty years of production AI in financial services, healthcare, retail, and enterprise technology is that decisions are only as good as the data beneath them. That is why I build provenance and traceability into agentic systems from the first design, not as an afterthought. I've applied the same discipline at scale before AI. For example at Cradlepoint, where my team took an alerts subsystem from about two alerts per second to 10x capacity using Kafka and a Cassandra backend.