I’m a technical leader who builds language systems that ship — from RAG pipelines and AI agents to the documentation and governance scaffolding that makes them safe in production.
Currently focused on GenAI content strategy after a decade across Pinecone, Splunk, and Workday. I move fast on prototypes, slow on the parts that matter, and care about the gap between a demo and a deployed system.
Outside of work I write essays on the limits of language models and build small tools for student ministry contexts.
LeCun, Lewis, and the Limits of Language ↗
An essay on hallucination, knowledge graphs, and what Yann LeCun gets right (and what C.S. Lewis still has to say) about the ceiling of language models.
Pinecone Search Documentation ↗
Owned the SDK, API, and retrieval narrative for Pinecone Search — the documentation surface developers hit first when evaluating vector retrieval at scale.
Project SAGE ↗
Selection Agent with Guided Explanations — a RAG-based support agent that synthesizes live DOM signals with static documentation to resolve tickets instantly. Top hackathon winner.
DocOps Governance Workbench ↗
A LangChain + MCP workbench for documentation governance — surfacing drift, policy violations, and freshness signals across large doc estates.
YouTube Content Pipeline ↗
An end-to-end content pipeline — Modal + PyTorch + ffmpeg + Claude — that converts long-form video into structured, searchable artifacts.
Splunk IX AI Council ↗
Founded a cross-functional governance body to operationalize Generative AI strategy across Splunk's information experience org.
QuickRamp ↗
A RAG-powered onboarding companion for healthcare clinicians, built on Next.js, Claude, and Pinecone with HIPAA-aware retrieval boundaries.
Fold ↗
A student-ministry tooling experiment — Next.js, Claude API, Drizzle, and a strong opinion that not every product needs an enterprise SaaS shape.