Raleigh and Durham form the Research Triangle, known for biotechnology and pharmaceuticals, for clinical research, for its universities, and for software and networking. Much of that work happens in laboratories and in studies, where a wrong answer is expensive and an invented one is worse. This page describes, sector by sector, what an AI engineer in Mexico can do for a company of that kind and where he stops.
Laboratories and the documents they run on
A laboratory keeps what it knows in protocols, methods, notebooks and reports. People in biotechnology and pharmaceutical companies tend to want one thing from AI before any other: to put a question to that material and receive an answer that names its source, or an honest "not found". That work is described under RAG and data platforms.
No case study on this site comes from biotechnology or pharmaceuticals. What transfers is a system of the same shape in another field. For a lean manufacturing practice I built an open corpus of documentation and an assistant that answers from it alone, and the case study shows how the search behind it combines meaning and exact words. A laboratory's corpus is private and controlled. That changes who may read what, and it does not change the method.
Clinical research
Organizations that run studies answer for every step: who changed a record, when, and on what basis. An agent in that setting has to leave the same kind of trail, and a change to it has to pass tests before it reaches anyone. No case study on this site comes from clinical research. One does come from a laboratory of another kind: a COVID-19 testing platform, with a role for each operator and a confirmation at every stage from registration to result. It is operations software that I built in 2021 and 2022, and it is not an AI system. What transfers is the guide on agent auditability, which sets out the evidence an agent should leave, and the release gates described further down. I do not interpret the rules on validated systems or on patient data. Your quality team writes the requirement, and I build to it.
Software, networking and the universities
Software and networking companies usually have an agent in a product already. Their question is whether tomorrow's version can ship. That is LLM evaluation, the center of my published work: a short set of cases on every merge, the full regression before a version is promoted, and each case run as its own synthetic user so that one test cannot leak into the next. My open-source MCP servers and my research on agents are public for the same readers.
Universities are the one sector here with a close match. Those gates belong to a multi-agent system that serves a higher-education group with 700,000+ students. It answers questions about admissions and enrollment, schedules appointments and sends each person to the right human team.
Eastern time against Monterrey's
Monterrey does not change its clock. From early November to mid March, Raleigh and Durham are one hour ahead of Monterrey. From mid March to early November they are two hours ahead, on the same Eastern hour as New York. I work inside your repositories and your cloud accounts, and review happens on the code and in writing, with a call when a decision calls for one.
Before you write
Tell me the role or the problem. I will read how the work is done now and reply with a first piece to build, the access it requires and the things I would not promise. I take senior roles and contract engagements. I am not a person in the Triangle and not an authority on your regulation. If the work needs either one, this is the wrong page.