Automation and AI-integration work, built like production software.

I design and build automated systems for businesses that need software to keep working when no one's watching it — webhook pipelines, LLM-backed workflows, and the n8n orchestration around them. 14 years of enterprise engineering is the standard I hold that work to, not a credential I'm bolting onto an AI wrapper.

What that discipline looks like in practice

• True idempotency — if a webhook fires three times because of a network stutter, your data doesn't get triplicated.
• Defensive error routing — if an LLM returns a malformed payload, it's quarantined and flagged, not silently dropped or forwarded broken.
• Clean data modeling — the relational schema behind your automated tools feeds your actual reporting, instead of fighting it.

Who this is for

B2B operators and small teams who are already making sales or running the business, and need the manual, error-prone parts — intake, routing, follow-up, reporting — automated without breaking under real volume. I've done this work for a real-estate brokerage and for a local food business; the underlying problem (data getting lost between systems that don't talk to each other) isn't specific to either vertical.

See me in action

The toolkit

• Languages & logic: TypeScript, Node.js, Python, JSON Schema validation
• Automation & flow: n8n, custom webhooks, message queues (BullMQ/Redis)
• Data & AI: PostgreSQL, Supabase, LLM integration with structured outputs
• Infrastructure: Docker, Cloudflare, Git-versioned deploys