Nine years at the same agency, five different jobs.
Open to AI Engineer and Applied AI roles · Mexico City, hybrid or remoteBefore EPA I worked in below-the-line marketing, and it bugged me that nobody knew whether anything worked: things kept happening because they always had. I wanted something I could measure, so I moved to digital. I joined EPA.Digital in 2017 as an account executive.
The first thing I automated was my reporting, in what was then called Google Data Studio, because I hated repeating myself in every presentation. Then I tried to automate insertion orders and ad uploads and couldn’t. It was way beyond what I knew how to do, and that drove me crazy.
One morning a PPC Hero newsletter described someone using R and a linear regression to reallocate budget. I downloaded RStudio thinking it would be like AdWords Editor, and two minutes later realized it was a programming language with statistics built in. I learned it in the afternoons. The agency’s CEO kept telling me I was doing it the hardest way possible; I didn’t listen.
That taste for the hard way cost me too. I showed up to client meetings with correlation matrices nobody could read, and had to learn to explain data in the client’s terms. Years later, when ChatGPT came out, I first used it for something I’d failed to solve with NLP models: ad copy. Then I saw it could also write code, and that’s what I’ve done since.
Today I’m AI & Automation Lead at EPA.Digital, an agency of about 200 people. I lead one AI developer, I’m product owner of internal products such as Einstein, Kalman and the Prompt Library, and technical co-owner of the media MCP server account teams use. On my own time I build Iterum, Domi and Vocatio.
I build them after work, in Claude Code and Claude Design. The part I enjoy most is design: bouncing ideas around to decide what goes in and what doesn’t, always on two pillars, AI at the core and the user experience. That’s where I learned something EPA didn’t teach me: shipping an app end to end without coordinating with another team, as its first user, and then testing it with beta users.
What I’m still working on: order when I build (that’s why I made Fundamentum), slowing down when I explain long ideas, and my English, which sits between B1 and B2.
Outside work I like to travel, around Mexico and abroad. I lived in Baja California Sur for a while and I’m a big fan of road trips and off-road driving. I watch a lot of HBO, like Succession and The Wire, The Simpsons, and all the science fiction I can.
All at EPA.Digital, since 2017
EPA.Digital · Mexico City · July 2017 – today
I lead AI adoption at the agency: I build and run the systems account teams use, lead one AI developer and advise Development, Measurement, Data Science and UX. I work with the CEO’s office on its “AI first” initiatives.
- Media MCP server: I built the first version in 2025 and co-own the production one, an agentic assistant over Meta, Google Ads, GA4, TikTok, Bing and DV360 on Cloud Run. I led the quality review of 484 real conversations.
- Prompt Library: I designed and built it (Next.js, Firestore, Cloud Run) and it opened to all 170 people at EPA, with live usage analytics and 0 USD a month to run.
- Product owner of Einstein, the agency’s RAG knowledge base with an MCP connector, and lead of Kalman, the AI-first reporting platform.
- Hopper: a QA portal for dashboards that replaced a 993-row spreadsheet with a guided flow that opens prioritized GitHub issues.
- SKU dashboard and ETL for Coppel, co-developed: daily analysis went from a full day to 30 minutes.
- Vibecoding governance: the agency’s Claude Code plugin marketplace, with a fixed stack, security guardrails and 35 media skills.
- Internal courses on prompting, AI Fluency and “Dashboards with Claude Code”, plus a GEO course at Universidad Panamericana.
Owned the data area: the central database, automated reporting, client analyses and internal tools.
- Designed the agency’s media database on BigQuery and an in-house Supermetrics replacement, the origin of today’s data platform.
- Benchmarks by channel and campaign type, and Slack bots with daily KPIs for e-commerce accounts.
- Standardized client analyses in R and Quarto: mobile UX audits, price competitiveness and Performance Max product clustering.
- Media automation: Google Ads scripts and a keyword optimizer with human approval in Sheets.
- Explored Marketing Mix Modeling with Robyn and forecasting with Prophet.
Created and led the conversion optimization program with a team of three (UX, development and analysis). Later I worked with the Marketing Director on CRO and data analysis: market KPIs and automated reports with R and Google Sheets. That led to the data manager role.
Analysis with Google Analytics and R, which I taught myself, Google Ads scripts and report automation.
I handled client relationships for campaigns on Google, Facebook and Bing: calls, tweaks, ad copy with its rounds of revisions and insertion orders done by hand. That’s where I automated my first reports in Google Data Studio.
What I work with
Only what a project backs up.
AI and agents
- MCP
- Claude · Anthropic SDK
- Claude Code
- Prompt engineering
- Gemini · Vertex AI · ADK
- LangGraph
- RAG
- Evals · guardrails
- OpenAI SDK
Data
- SQL · BigQuery
- ETL
- Python
- Looker Studio
- Firestore
- GA4 · Google Ads · Meta · TikTok · DV360
- R · Quarto
- PostgreSQL
Engineering and cloud
- TypeScript · Node
- Next.js · React
- Cloud Run
- Cloud Functions · Scheduler
- Apps Script
- Terraform
- pytest · vitest
Education, certifications and languages
- Degree
- B.A. in Business Administration · UNAM, School of Accounting and Administration · graduated 2016
- Certifications
- Launching into Machine Learning · Google · 2024
- Prompt Design in Vertex AI · Google · 2024
- Data Analyst with Python · DataCamp · 2022
- Languages
- Spanish · native
- English · intermediate (B1–B2), taking classes
Shall we talk?
I’m looking for an AI Engineer or Applied AI role in Mexico City, hybrid or remote, with a proactive team that moves without waiting to be told what to do. After five different jobs, I want to specialize in AI and agents.