Four ways to work together.
Pick the one that fits where you are. In all of them you get the same thing: a designer who builds, and one person who answers for the result.
- Remote from Paraguay · UTC-3
- US teams since 2021
- English and Spanish
Build your product
For founders and companies starting a product
From the database schema to production: backend, frontend, infrastructure and deploy. One engineer, end to end, so nothing gets lost between design, API and server.
What’s included
- Product design first: flows and screens you approve
- SaaS or web app, multi-tenant when you need it
- Auth, roles, payments, queues and integrations
- Your own infrastructure, CI and monitoring
- Handover docs, or I keep running it
Join your team
For product teams that need senior capacity
A senior full-stack or frontend engineer working inside your codebase and your process, on your hours. Long engagements, not hand-offs.
What’s included
- Full-stack or frontend, React and TypeScript first
- Works in your repo, tickets, reviews and CI
- Owns features end to end, tests included
- Full overlap with US Eastern (UTC-3)
- Engagements of 10 to 23 months so far
Design to code
For teams without a designer, or with design debt
Figma UI and UX, design systems shipped as packages, and mobile apps, built by the same person who designed them. Nothing gets lost between Figma and production.
What’s included
- UI/UX in Figma: flows, screens, prototypes
- Design systems: tokens, components, Storybook
- Web and mobile (React Native) implementation
- Accessible by default: WCAG 2.2 AA
- Gallery of past work on request
<Button variant="ink" href="/contact/"> Start a project</Button>AI & automation
For companies with manual work, or that want AI features
LLM features that keep working when a provider fails, integrations between the tools you already use, and bots that take repetitive work off your team.
What’s included
- LLM features with provider fallback and cost control
- Document and invoice extraction bots
- Integrations: Stripe, HubSpot, Bancard, Moodle, Okta, ERPs
- Local models when data can’t leave your servers
- Logs and retries you can audit
“Forecast tomorrow’s queue”
Provider routertimeouts · retries · cost
- Gemini Flash Liteprimary
- Claude Haikufallback
- Local modelprivate
How a project runs
Same steps, whichever service you pick.
Design first
Screens and flows before code, so you approve something you can see.
Small slices
Production deploys from the first weeks, CI on every change.
One person accountable
Schema, API, UI and infra. No hand-offs between people.
Your hours
UTC-3, full overlap with US Eastern.
No fixed packages
Every project is scoped from a short call; the form asks for a budget range.
Reply in 24 h
Tell me what you’re building and I answer within one business day.
EN · ES
Calls, docs and code reviews in English or Spanish.
Not sure which one fits?
Describe the problem. I’ll tell you how I’d approach it.