Product leader with 10+ years building and scaling high-volume digital platforms across LATAM (Rappi, Mercado Libre, Despegar). I work at the intersection of pricing, monetization, marketplaces, payments/fintech and platform/API integrations, and I'm now focused on rebuilding product and teams around AI.
📍 Buenos Aires, Argentina
Leading the shift to AI-native product: rethinking how product gets built and how teams operate around agents (Claude Code, Cowork, Cursor, Codex), not "AI bolted on."
The working thesis is that AI-native means building your own tools. So instead of writing about it, I run my work and my week on a system I build in the open. Everything below is the real thing with the personal data stripped out.
| Repo | What it is |
|---|---|
| personal-os | An AI-native personal operating system: a router agent that opens every session, four domain copilots that own their territory, a catalog of skills they invoke, and a memory loop that makes it improve with use. |
| pm-skills | Reusable skills for PM work: PRDs, metrics reviews, research synthesis. Evals-first. |
| decision-canvas | Turns an analysis into an interactive decision artifact: compare alternatives, inspect assumptions, test sensitivities before committing. |
| send-to-kindle | Markdown to reflowable EPUB3 to your Kindle, so long deliverables get read away from a screen. |
| md-to-podcast | Turns a document into a full two-voice audio walkthrough for the car. Full coverage, not a summary. |
The pattern I keep coming back to: write the evals before the code. Every skill above ships with a behavior spec, an eval set built from real failures, and an autonomy map that says what the agent does on its own and what it has to ask about. personal-os documents the whole architecture in ARCHITECTURE.md.
- Pricing as demand-shaping. Fees, slots and incentives as one system to move conversion, affordability and unit economics, not just a revenue lever.
- Marketplace and merchant economics. Monetization, integrations and the flywheel between supply, demand and margin.
- AI-native product development. Designing products and workflows around new capabilities from the ground up. I'm especially into closing the loop from user feedback + data → synthesis → prototype: systems that compile raw user signal and turn it into working solutions built with AI.
- Self-improving agents. An agent that evaluates itself before answering and persists what it learns afterwards gets better every week. That loop is the interesting part, not the prompt.
- Rappi, Head of Product (35M+ users across 9 LATAM markets). Pricing & Checkout, and Restaurants Integrations. Owned the end-to-end mobile funnel with a focus on growth, monetization and EBITDA. Scaled integrated-order share 35% → 60% and drove eight-figure annual EBITDA impact through a demand-shaping approach to pricing.
- Despegar, Group Product Manager. M&A Integrations & B2B platforms. Led a team of PMs and grew B2B GMV contribution <2% → 35% via white-label platforms for banks, retailers and travel companies.
- Mercado Libre (Mercado Pago), UX Product Specialist. Founding team member for in-store QR payments; also shaped wallet, investments, asset-management and unified checkout experiences across mobile and web.
Crehana, Professor of Product Management (2020-2023). Built and taught Growth & Monetization in a PM certification validated by UTEC (Peru): pricing, revenue management, conversion funnel, CAC/LTV and product strategy.