> kyle_dumipnas --work --stack --about --contact
Book a call

Kyle Andrew Dumipnas

AI Automation Engineer & SaaS Developer

I design and build n8n and GoHighLevel automations, and the backend services that run behind them — Python, FastAPI, and Node.js. Real builds, demoed on video, not slideware.

Book a call See the work

Or find me on GitHub · LinkedIn

automation.flow
GHL New lead contact created n8n Webhook trigger receives the payload n8n Filter intent = support n8n Score priority = high GHL Update contact ✓ synced to CRM

Built with

Python
FastAPI
Node.js
n8n
GoHighLevel
Docker
GitHub
Supabase
PostgreSQL

Selected work

Real builds, not demos — 4 automations shipped in n8n and GoHighLevel.

Loom walkthrough coming soon

Add your video URL in the loomUrl field

AI-Powered Email Triage

Incoming emails are auto-classified by intent, priority-flagged, and routed to the right inbox or owner — no manual sorting.

Client
—
Timeline
—
Complexity
Medium
Impact
Removes manual sorting from the support inbox — incoming email is triaged and routed automatically instead of by hand.

Workflow steps

  1. 1 New email arrives via inbox trigger (Gmail/IMAP).
  2. 2 AI model classifies intent and urgency.
  3. 3 Priority flag is attached to the message.
  4. 4 Message is routed to the correct inbox or owner.

Loom walkthrough coming soon

Add your video URL in the loomUrl field

Workflow Version Control to GitHub

n8n workflows are auto-synced to a GitHub repo on every change, giving every automation a real version history and a rollback path.

Client
—
Timeline
—
Complexity
Medium
Impact
Every workflow change is versioned and reviewable, with a real rollback path instead of relying on n8n's built-in history.

Workflow steps

  1. 1 n8n workflow is saved or updated.
  2. 2 Change is detected and exported as JSON.
  3. 3 Export is committed to a GitHub repo.
  4. 4 Full version history and diffs become available in Git.

Loom walkthrough coming soon

Add your video URL in the loomUrl field

File Ingestion + RAG + Auto CRM

Incoming documents are parsed and processed with a retrieval-augmented AI pipeline, then used to auto-populate CRM records — no manual data entry.

Client
—
Timeline
—
Complexity
High
Impact
Eliminates manual data entry from incoming documents into the CRM.

Workflow steps

  1. 1 Document arrives (upload or inbox trigger).
  2. 2 Text is parsed and chunked for retrieval.
  3. 3 Retrieval-augmented model extracts the relevant fields.
  4. 4 Extracted fields auto-populate the CRM record.

Loom walkthrough coming soon

Add your video URL in the loomUrl field

AI Lead Qualifier

New GoHighLevel leads are automatically scored and qualified by AI before they reach a human — only real opportunities get a follow-up call.

Client
—
Timeline
—
Complexity
Medium
Impact
Only qualified opportunities reach a human for a follow-up call.

Workflow steps

  1. 1 New lead is created in GoHighLevel.
  2. 2 Lead details are scored by an AI model.
  3. 3 Qualified leads are flagged for follow-up.
  4. 4 Unqualified leads are routed to nurture instead of a call.

How an automation gets built

Same process on every project, whether it's n8n or GoHighLevel underneath.

  1. 01

    Map the workflow

    Walk the manual process end to end — the tools, the handoffs, the steps someone repeats every day.

  2. 02

    Build the automation

    Triggers, logic, and error handling in n8n or GoHighLevel — version-controlled, not a black box.

  3. 03

    Ship and hand off

    Runs in production, documented, with a clear owner for changes after I'm gone.

About

whoami
Kyle Andrew Dumipnas

I build automation systems on n8n and GoHighLevel — the kind that quietly do the repetitive work so people don't have to. Most of what's below started as a manual process someone was tired of repeating: sorting email, copying data between tools, tracking down which workflow changed and when.

When the automation needs more than a workflow canvas can give it — a real API, a database, a scheduled job — I build that layer too, in Python, FastAPI, or Node.js, containerized with Docker and backed by Supabase or SQL.

Each project on this page is a real build, not a demo. If something here is close to a problem you're carrying, that's a good place to start a conversation.

Let's build your next automation.

Tell me what's eating your team's time — I'll tell you whether it's a good candidate for automation before we get on a call.

Book a call Email instead