Process automation, no fluff.
I spent years moving between large companies like Microsoft and early-stage startups. On both sides I kept seeing the exact same pattern: back-office work eats most of a team's time, and almost all of it is repetitive. People don't fall short because they lack judgment. They fall short because they're handling tasks the tooling should be doing on its own.
The fastest way to get a worker (and a business) operating at 100% is removing the friction between the tools they already use and the outcomes the company actually needs.
That's where Data Quimbaya comes in: we study your data so we can act on it, not just report it. We design the flow, connect the pieces, and leave the human deciding what only a human should decide.
Human-in-the-loop by default
AI accelerates, it doesn't replace judgment. I design flows where the critical final call always stays with a person.
Solutions, not platforms
If an existing tool solves the problem, we use it. If not, I build just what's needed, no over-engineering.
Real measurement, not promises
Every automation ships with metrics: hours saved, conversion lifted, errors reduced. If it doesn't move a number, we don't charge for it.
Technical stack
I'm fluent in:
- No-code platforms: n8n (preferred), Make, Zapier
- Languages: TypeScript/Node, Python
- AI: Claude API, OpenAI API, embeddings, RAG
- CRMs and infra: HubSpot, Pipedrive, basic Salesforce, Postgres, Supabase
- Messaging and productivity: Slack, WhatsApp Business API, Google Workspace