Technology

As a full-stack JavaScript/TypeScript developer I work primarily in the React, Next.js, and Node ecosystems. I build and maintain UI architecture, CMS/content modeling, API integration, performance, and deployment— stepping into large, established codebases to ship improvements, and helping small scale clients with burgeoning web presence; I enjoy building experiences for working creatives, where design and implementation evolve together.

The most complete expression of that work so far is my agentic chatbot mixer — a multi-model AI group chat where a server-side orchestrator plans every conversational turn through typed structured output: which models speak, what gets relayed between private threads and the group, when someone reaches out unprompted. Building it meant solving concurrency (four simultaneous SSE streams with sequenced reveal timing), durable state (server-side session persistence with the client store as a render cache), and cost discipline (hard usage caps, because an unbounded personal project isn't shipped). It runs live on this site.

After the pandemic, I found time to teach and mentor young adults and kids through The Coding Space, where I saw a handful of students catch the programming bug, finding lifelong inspiration to build what they imagine, because they can.

With artificial intelligence rapidly reshaping our fields and roles, I work with AI from both directions: building agentic systems where models plan and act, and using tools like Claude Code and Cursor daily as multipliers on my own output. That second part comes with a responsibility I take seriously — the skill isn't prompting, it's knowing when generated code is confidently wrong. Efficient developers should continually audit new tools and strike a balance between what should be handwritten and what can — and should — be automated. Recently that auditing instinct has extended to infrastructure: containerizing production Next.js applications with Docker and Docker Compose.