I design systems.
I govern quality at scale.

The practice

My career has been a systematic accumulation of architectural perspective. I started as an automation engineer at UNAM — writing Selenium scripts, building APIs, operating Linux servers. Each subsequent role expanded the scope: quality governance at Keywords Studios, regression strategy at Wipro, fintech test architecture at Applica, LLM safety evaluation at Zurich Insurance. At Clip, the work is explicitly architectural — evaluating the feasibility and ROI of engineering initiatives alongside Architecture Design and Business units, governing quality across distributed teams in China and Argentina, and designing data integrity frameworks for high-volume fintech transaction systems.

IT Architecture is not a destination I am approaching — it is the discipline I practice. The questions that define it are the same ones that have driven every role I've held: What are the failure modes? What trade-offs are we making, and are they intentional? Where does this system break under load, under change, or under misuse? I ask those questions before the first line of code is written, and I build the governance structures that ensure the answers remain auditable as systems scale.

On accessibility

I grew up with a Deaf mother. Communication, in our house, was intentional — you learned to think about how information travels, who receives it, and what gets lost between sender and receiver. That shaped how I approach software.

Accessibility is not a compliance checkbox for me. It's a systems constraint — one that, when you design for it from the start, makes everything else more robust. The interfaces that work for someone using a screen reader or navigating without a mouse tend to be the same ones that load fast, scale cleanly, and don't break when conditions aren't ideal. Inclusive design and good architecture are not in tension. They're the same discipline, approached with the same rigor.

Current work

POLQA — an open-source framework for governing LLM output quality and evaluating political bias at scale. The design choice was deliberate: a modular CLI architecture over a monolithic evaluation suite, so teams can compose evaluation pipelines from curated datasets and prompt variants without being locked into a single provider. Reproducibility and auditability are first-class constraints, not afterthoughts.

This site — a professional record structured around how I think, not just what I've done. The writing is the signal. If you've read this far, you've already seen the approach.

Outside the terminal

I write music — I've been a songwriter long enough to know that the skills transfer: structure, economy of expression, knowing what to leave out. I'm based in Mexico City, and I work in three languages — Spanish (native), English (C1), and French (B1).