Monterrey is known for heavy manufacturing: steel, cement, glass and beverages. It is also known for the headquarters of industrial groups, for the suppliers that serve the relocation of plants, and for its universities. I live and work here. This page says what companies of those industries tend to need from AI, which published work of mine is rooted in the city, and how the work runs for a company that is somewhere else.
What a plant needs
A plant keeps its knowledge in procedures, in quality records and in the memory of the people who have run a line for years. The usual first request is to make that knowledge answerable at the moment it is needed. In practice that is an assistant that answers only from the plant's own documents and shows where each answer comes from, which is the work of RAG and data platforms. The second request is often smaller: a tool for one job on the floor, such as a calculation that lives in a spreadsheet today or a checklist that lives on paper.
What an industrial group needs
A group with many companies asks about control before it asks about models. Who may let an agent reach which system? Who signs when a version changes? What is recorded about each run? This is enterprise AI: narrow access to corporate data through an MCP gateway, evaluation gates that decide what ships, and an audit record of what the agent did. The guides on agent auditability and on enterprise agent architecture describe each mechanism.
The published work behind that answer is a multi-agent system that serves a higher-education group with 700,000+ students, where a version that fails its tests does not ship. It is also the closest published work for a university.
What a supplier needs
A supplier to relocating plants sells by quote. The time of its sales people goes into reading requests, preparing quotes and following up on them. The published work that transfers is a set of agents around a quoting and deal-tracking system for small and medium companies. The agent drafts the quote and updates the pipeline, and the seller decides the price. A supplier that finds its customers by email can also read how I treat deliverability and prospecting as engineering.
Work rooted in the city
A case study on this site comes from Monterrey: the platform of a lean manufacturing practice that works with food, metalworking and automotive plants. I built its open documentation corpus, its working tools (a maturity self-assessment and an OEE calculator), an assistant that answers from that corpus, and an MCP server so that an AI agent can read the same material a person reads.
The case shows how I treat manufacturing knowledge: written down, searchable, in Spanish and quoted with its source. It is the platform of a practice that works with plants. It is not a system installed on a production line, and I do not present it as one.
The clock
Monterrey is on UTC-6 all year, the same hour as the rest of central Mexico. Cities on US Central Time are on the same hour from early November to mid March and one hour ahead from mid March to early November. A company here that also works with Texas does not gain a third clock by working with me.
Companies elsewhere
I also work remotely with companies in other cities of Mexico and in the United States. The arrangement is the same as for a company here: the system is built in the company's repositories and cloud accounts, and the process is described by the people who run it.
How to start
Write to me with the role or the problem. I am open to a full-time senior role and to contract engagements under my own name. For contract work the first step is to read the process, and my first answer is in writing: what to build first and what to leave for later.