Ten Lives · festival, Portugal
Supplier scraping and enrichment pipeline
Across Portuguese directories, deduplicated and filterable by region, certification and category.
PythonGoogle Places API
Automation & data pipelines
Stop re-keying the same data into three tools.
When systems don’t talk to each other, people become the integration: copying rows, forwarding emails, checking one list against another. I build n8n and Python workflows that do that part, and check what they produce against the data you already trust.
Sound familiar?
When people are the integration.
Most of these don’t need new software. They need the tools you already pay for to pass data to each other.
- Someone copies data from one system into another every week.
- Incoming requests are read, sorted and forwarded by hand.
- A list gets compiled by hand and is out of date within a month.
- Onboarding someone means typing the same details into four forms.
- A workflow exists, but only in one person’s head.
What I build
Small workflows that fail loudly.
An automation nobody watches is worse than the manual step it replaced, because it goes wrong quietly. Everything I build says what it did.
Integrations
n8n workflows between the tools you already use, over their APIs and webhooks. No new platform to adopt.
Data pipelines
Python jobs that collect, clean, deduplicate and load data where it’s needed, on a schedule.
Triage and qualification
Incoming leads, requests or documents sorted by rules, with an AI step only where it earns its place.
Guardrails
Each step logs what it did, stops on bad input rather than passing it on, and can be switched off without breaking anything else.
Example: Ten Lives supplier pipeline
From scattered listings to one usable list.
Ten Lives is a regenerative community festival in Portugal. The supplier information it needed was spread across Portuguese directories.
- 1 · CollectDirectories and listings
Supplier data gathered across Portuguese directories with Python and the Google Places API.
- 2 · CleanDeduplicated
The same business listed in several places becomes one record.
- 3 · EnrichStructured
Each supplier tagged so the list can be filtered by region, certification and category.
- 4 · UseA list the team works from
One deduplicated list, filtered to what’s needed.
Recent automation work
Where I’ve done this.
Ten Lives · festival, Portugal
In progress
Volunteer onboarding pipeline
An onboarding pipeline for the festival’s volunteers, for the same client. Still being built, so there is no outcome to report yet.
Rapidos.ai · contract, 2025–26
Lead qualification and knowledge bases
Designed automated knowledge bases and lead-qualification flows, and translated business requirements into n8n workflows and internal AI agents.
n8nn8n AI agents
Galene Pathways · co-founder, 2023–25
Environmental data pipelines
Built internal data workflows and scripts for a marine technology startup, and ran user discovery to test the data assumptions and the human-in-the-loop steps.
PythonData workflows
Stack
What I work in.
Automation
n8nn8n AI agentsREST / webhooks
MCP servers
Data
PythonPandasSQL / PostgreSQL
SupabaseCSV / XLSX ingestion
Delivery
AI-assisted development (Claude Code)Git / GitHub Actions
Docker
If the question is less “how do I automate this” and more “should AI be involved at all, and with which data”, that’s advisory.
How it runs
One workflow first, on real data.
Then the next one, if the first one earned it.
- First step
- A 30-minute call about one manual step. Free.
- Shape
- One workflow built end to end and checked against real data, before the next one starts.
- Where it runs
- On your accounts, with your credentials. I don’t host your automations.
- Price
- Quoted after a call.
- Handover
- Each workflow documented: what triggers it, what it touches, and how to switch it off. Support afterwards if you want it; nothing depends on it.
Describe one manual step.
What gets copied, from where, to where, and how often. I’ll tell you whether it’s worth automating, and what that would involve.