AI Engineering

Agents, knowledge bases, and tooling — built into a real product, not bolted on.

I help teams and founders bring AI into actual workflows — not as a chatbot demo, but as a reliable layer that ships work. Multi-agent orchestration, custom skills/tools, knowledge bases, prompt design, and pragmatic token economics.

What I deliver

Multi-agent orchestration: Team Lead / Frontend / Backend / Logic Operator setups (built before mainstream orchestrators were public). Roles, hand-offs, validation gates.
Knowledge bases: NotebookLM for fast personal research; self-hosted MD-repo "super-agents" when you need control over content, deep analytics, or Mind Maps.
Custom skills, tools, automations: scripts that batch and structure work instead of round-tripping through chat — keeps token spend tight without sacrificing quality.
Prompt engineering in production: not one-off prompts, but versioned, evaluated, and tied to clear behavioral specs.
Tooling stack: Claude (Code, Skills, Brainstorming, Superpowers), ChatGPT, Gemini, Codex, Grok, Bolt.new, v0.dev, Magic Patterns, Aura Build, Google Stitch, DALL-E, Nano Banana.
SEO + AI: GEO/AEO strategy — how brands get cited inside Claude/ChatGPT/Perplexity answers, not just ranked on Google.
Custom developer tools: e.g. an in-house SEO crawler that surfaces site issues better than off-the-shelf browser extensions.