Portfolio
I've spent 20+ years building products across very different domains: secure messaging, banking, e-commerce, tax compliance, civic tech. Sometimes as an engineer, sometimes as a product manager, often as both at once. This portfolio collects that work, from professional products to hackathon experiments. The featured ones follow the same logic: the problem, what I did, and what came out of it.
Featured Case Studies
Transparencia ES: public procurement, explained to anyone
2026 Claude Agent Skill + conversational web app Civic tech, open data
The problem. In Spain you have the right to know how public money is spent, but the data is scattered across a dozen portals behind expert jargon: CPV codes, contract types, tender IDs. Civio's journalists carry that knowledge in their heads; a normal citizen doesn't.
What I did. Built at Ship for Good, the Software Crafters Barcelona hackathon for Fundación Civio, with Mercedes Egido. We designed two layers: a Claude Agent Skill as the brain, systematizing the journalists' tacit knowledge, translating jargon into plain language and open to community-contributed parsers for more transparency portals; and a conversational web UI that never invents and always links the official source.
The outcome. A working MVP in a weekend: the web runs on cached Zaragoza city council data, the skill queries live PLACSP data. The next step is what matters most: going from output to outcome, giving it real life beyond the hackathon.
DemoRepoSkill
Permalink to this case study →
Quaderno: tax registration and filing in 50+ countries
2025 Technical Product Manager Bootstrapped SaaS, ~15 people
The problem. Quaderno covered tax calculation and reporting, but not registration and filing: the part of the tax lifecycle where customers still had to go elsewhere. Sales were declining and we needed more LTV per customer. Filing requires precise manual work, per-jurisdiction expertise and legal responsibility, so operating it in-house was not viable.
What I did. I proposed and validated the hypothesis of working with partners whose core business is filing. We started with manual per-country operation to validate the learnings before building. Then I designed the partner integration model (APIs, data flows, error handling) and the abstractions that captured what countries have in common while isolating what is specific, so adding a country became configuration instead of development, and implemented the integrations myself in the Rails codebase. I acted as the bridge between Sales, C-level and the external partners.
The outcome. In 8 months: hundreds of new tax registrations across 50+ countries, 10x LTV among the customers who adopted it, and the team did not need to scale. Supporting a new market went from weeks to days. It wasn't perfect, though: there's always a country that breaks your assumptions.
RegistrationReporting & filingSupport guide I wrote
Permalink to this case study →
developers.quaderno.io & Quaderno Connect: docs as a product
2021 to 2025 Lead Technical Writer, Developer Relations Ruby on Rails, OpenAPI, rswag
The problem. Quaderno's API products needed self-service integration: with a ~15 people company, every integrator question that reached support was expensive, and outdated docs are worse than no docs. The stakes grew with Quaderno Connect (https://developers.quaderno.io/guides/connect/), the white-label API for platforms and marketplaces: complex integrations, a domain full of edge cases (taxes), and big prospects comparing us against alternatives like Stripe Tax.
What I did. I built (and wrote!) the whole developer site (https://developers.quaderno.io) and evolved the API, treating both as products: OpenAPI reference, integration guides per user profile. Particularly proud of making the OpenAPI reference auto-generate from tests in the Rails codebase (rswag), which gave ownership to the whole team: keeping docs up to date became the path of least resistance. For Connect I owned the product side: reference and guides, hands-on support to integrators through implementation (including the complex tax-domain explanations), evolving and extending the API from their needs directly in the codebase, and working with Marketing on the landings and with Sales on positioning.
The outcome. Reduced integrator questions and became a huge selling point over competitors: large international marketplaces evaluating alternatives like Stripe Tax chose Quaderno, and they repeatedly credited the documentation and the implementation support. I gave a talk about this approach at Software Crafters Barcelona 2025, "Documenting APIs" (https://www.youtube.com/watch?v=IHIsuIlSiio), and I'm bringing an updated version to Write The Docs Berlin in September 2026.
Permalink to this case study →
Flywire: validating a new business line
2016 to 2017 Product Developer at Coding Stones Ruby, Vue.js, Docker, Terraform
The problem. Flywire had a solid product for student-to-university payments and wanted to explore a new line: university-to-supplier payments. Lots of uncertainty about what the business actually needed, and a real risk of spending months building something that didn't fit.
What I did. Instead of jumping to the full solution, we ran a Lean Inception with their team: identified users, mapped needs into a shared User Story Map, and prioritized an MVP that would validate the business case before committing the full budget. Each story went through Example Mapping into BDD acceptance tests, with an internal dashboard visualizing real progress from the tests.
The outcome. We delivered something smaller and rougher than the client initially expected. In exchange, they learned fast and adjusted before the big investment. Business case validated; Flywire continued the full development in-house, and later became the first Spanish startup to debut on the Nasdaq.
Permalink to this case study →