Senior AI engineer in Latin America (LATAM), from Mexico

I am a Senior AI Engineer in Mexico and I work remotely from Monterrey, in Spanish and in English. A company that searches for an AI engineer in Latin America finds me in one country of the region, and this page says what that means from the others.

A search for an AI engineer in Latin America usually comes from one of two places. A company in the region wants a senior engineer who works in its language and close to its clock. A company outside the region wants one engineer who can work with its operations in several countries. I am in Mexico, in Monterrey, and the work is remote. This page says what that means from Bogotá, Lima, Santiago, Buenos Aires or São Paulo, which published work touches the region, and what I do not claim.

The clock from your city

Monterrey is on UTC-6 all year. Against that fixed hour:

  • Bogotá and Lima are one hour ahead of Monterrey all year.
  • Santiago is two or three hours ahead, depending on the season.
  • Buenos Aires and São Paulo are three hours ahead all year.

With Bogotá or Lima almost the whole working day is shared. With Buenos Aires or São Paulo the shared part is shorter, and calls go in the hours when it is morning in Monterrey and already afternoon there.

The language of the process

I work in Spanish and in English. For a company whose people work in Spanish, that takes translation out of the project. The person who knows how an order, a quote or a complaint is handled explains it in their own words. The cases that test an agent are written the way its users write, with their terms and their abbreviations. The document that says what the agent may and may not do can be read by the person who has to approve it. With a team in São Paulo the work would run in English or in Spanish, the two languages I work in.

Published work that touches the region

A quoting system for small and medium companies in Latin America. The case study describes agents around a quoting and deal-tracking system. Its starting point is that a quote is a decision in motion, not a document. The agent carries the path that repeats around every quote: it reads the request, looks up the product and the price, drafts the quote, prepares the follow-up message for WhatsApp and updates the pipeline. The seller keeps the judgment on price and the relationship with the customer. Each run leaves a trace that a person can review.

The site in Spanish. The services, the engineering guides and the CV are also published in Spanish. The Spanish version of this page is the way in.

What companies tend to ask for

The request changes with the kind of company more than with the country.

A distributor or a wholesaler starts from the quote, because that is where the hours of its sales people go. The service is AI agent development, and the quoting case above is the closest published work.

A plant or one of its suppliers starts from knowledge that sits in documents and in senior people. What I can show is the platform of a lean manufacturing practice in Monterrey that works with food, metalworking and automotive plants: an open corpus, working tools, an assistant grounded in that corpus and an MCP server.

A corporate group, a bank or a university starts from control: who may let an agent reach which system, and who signs when a version changes. The published work is a set of evaluation gates in a multi-agent system that serves a higher-education group with 700,000+ students, described in its case study. No case study on this site comes from banking or from government. What transfers is the gate itself and the audit record of each run.

A software company asks for tests before its agent reaches users, or for an MCP server over its own product. The site shows open-source MCP servers, the evaluation of LLM systems and pre-registered research on agents.

What I do not claim

I am in one city of one country of the region. This page lists no client in any country. What it offers can be checked instead: the case studies, the public code and the research.

How to start

Write to me in Spanish or in English with the role or the problem. I am open to a full-time senior role and to contract work under my own name. My first answer is in writing: what I would build first, and what I would leave out.

Common questions

How do I find the best AI engineer in Latin America?

No list can answer that for your company. What can be checked is the work: a system in production, the evaluation that blocks a bad release, and an honest account of something that broke. The guide on choosing an AI engineer gives the questions, and they apply to me as much as to anyone else on your list.

Which country of Latin America are you in?

Mexico. I live in Monterrey and I am not in any other country of the region. With a company in Bogotá, Lima, Santiago, Buenos Aires or São Paulo the work is remote.

What is the time difference between Monterrey and Bogotá or Lima?

One hour, all year. Bogotá and Lima are one hour ahead of Monterrey, which stays on UTC-6 in every season. Almost the whole working day is shared.

What is the time difference with Buenos Aires, São Paulo and Santiago?

Buenos Aires and São Paulo are three hours ahead of Monterrey all year. Santiago is two or three hours ahead, depending on the season. Calls fit best in the hours when it is morning in Monterrey.

Which of your published work is closest to a company in Latin America?

The case study on agents around a quoting and deal-tracking system for small and medium companies in Latin America. The services, the guides and the CV are also published in Spanish. Neither is a statement about clients in a given country.

Open to senior AI engineering roles and contract engagements, remote in the US, Mexico and Latin America.