About
Software engineer in Rancagua, Chile. Five years building, four of them in systems where money moves.
How I got into this
I started Industrial Civil Engineering at Universidad de O'Higgins in 2018 and spent two years as a teaching assistant in Programming, Differential Equations and Physics. Teaching the programming course is what made me want to write software instead of studying industrial processes.
In 2021 I left the degree, spent three months studying full time, and got my first developer job at Tactech, where I built the menu and order management microservice for restaurants and integrated its delivery and payment platforms: Mercado Pago, Uber Eats, Rappi and PedidosYa.
Four years in fintech
At MACH I built the international transfers system from scratch, integrated with the banking core and with VISA, as microservices in Node.js, TypeScript and MongoDB. I optimized the savings and investment flows and implemented credit card debt payments from inside the app.
At Mercado Libre I worked in the Cash team on Mercado Pago's money flows for Chile, Argentina, Mexico and Brazil. I automated cash-in and cash-out point loading, which went from a week down to a few hours, and moved vulnerability management for the team's 40 repositories to an automatic flow. Money systems taught me the habit I care about most: shipping things that cannot fail quietly.
What I build now
I lead the technical agility of an engineering team at Sunix. I built the Sunix AI Agent Platform that runs the company's sales and support on web and WhatsApp with human on the loop, its MCP server that exposes the platform as tools scoped by role, its CMS and integration SDK, and the CI/CD behind all of it.
Two more products of mine run in production: NexusTML, a mobile app for industrial maintenance used in a plastics plant, and CuidaRed, a telemedicine platform for a network of health centers.
How I work
I am self-taught, so I learn by building and by reading source. I care about tests that describe behavior, about CI that catches problems before a person does, and about code the next engineer can change without fear. I use AI every day as part of the workflow, not as a replacement for understanding the system.