I build production AI automation systems that clients actually use every day. The kind that quietly run in the background and give a team back hours every week.
My core is n8n and AI. I design multi-step workflows with proper error handling, retries, and deduplication, and wire in the Claude and OpenAI APIs for extraction, scoring, classification, and content tasks.
I work in a modern AI-native style: I direct Claude Code (through Emdash, an agentic dev environment where I run it across parallel Git worktrees) to build and modify code, then review, test, and ship. Fast to prototype, careful about what actually holds up in production.
WHAT I DO
- Workflow automation (n8n): lead generation, AI data processing, email automation, and CRM syncs, with reliability built in
- AI agents and assistants: agents on the Claude and OpenAI APIs, plus AI voice agents (Vapi, Retell). Custom assistants and knowledge bases via Claude Projects, Claude Skills, and MCP servers
- API integration: REST, OAuth, and webhooks daily, connecting Google Workspace, Airtable, and other tools into clean data flows
- Prompt engineering: multi-step prompts and chains tuned for consistent, structured outputs
TOOLS
n8n, Claude and OpenAI APIs, Claude Code (via Emdash), Claude Projects and Skills, MCP servers, Vapi, Retell, REST/OAuth/Webhooks, JSON, Google Workspace (Gmail, Sheets, Drive), Airtable, Apify, Git/GitHub
I work independently, document what I build so it stays maintainable, and communicate clearly with international clients. The bar I aim for on every build: invisible enough that the team forgets it's there, reliable enough that they can.
Verified
