Writing

Ideas on product, AI systems, and building things.

Context as a Product Layer

How treating context as infrastructure — not a feature — changes everything downstream.

What it takes to give an AI real memory, identity awareness, and proactive intelligence.

The gap between having opinions about software and having built it yourself.

Small, well-placed architectural decisions that create leverage over time.

A practical introduction to vector search, semantic similarity, and why it matters for products.

Patterns for building AI agents that actually work — tool use, memory, and orchestration.