Remote AI engineer for companies in Philadelphia

I am a Senior AI Engineer and I work remotely from Monterrey, Mexico, with companies in Philadelphia, as a contractor or in a senior role. I am not in Pennsylvania: the work runs in your repositories and in writing, with Philadelphia one or two hours ahead of my clock.

Philadelphia sits between New York and Washington, and it is known for pharmaceuticals and life sciences, for its hospitals and universities, for financial services, and for its port and logistics. The first three are institutions that answer to patients, students and regulators. This page says what such an institution can expect from an AI engineer who works from Mexico.

Pharmaceutical companies

A pharmaceutical company is many departments with one habit: every claim is reviewed before it leaves the building. What these companies tend to ask of AI is help with the reading and the drafting inside that review, without a model inventing a figure or citing a document that does not exist. The service is Enterprise AI, and the part that matters most is the test a system must pass before a new version runs.

No case study on this site comes from pharmaceuticals or life sciences. What transfers is the discipline. In the evaluation work I published, anything a plain comparison can settle, such as an exact figure, is never left to a model's judgment. The model that judges the rest was checked against human labels, one case at a time, until the two agreed.

Hospitals and universities

Both receive a constant flow of questions from people who need the right desk, a date or a next step. That is the shape of the system in my case study on evaluation gates: agents that answer admissions and enrollment questions, schedule appointments and route each person to the right human team, for a higher-education group with 700,000+ students. For a university, that is work from its own sector.

For a hospital it is not. No case study on this site comes from a hospital. The nearest is a COVID-19 testing platform for a laboratory, which handled registrations, test certificates and results. I built it in 2021 and 2022, and it is operations software, not an AI system. What transfers is the structure of the problem (answer, schedule, hand over to a person) and the gates that keep a change from breaking it. The rules on patient data are yours to state, and I build inside them.

Financial services, the port and logistics

Financial firms ask what an agent did and who can prove it, and the guide on agent auditability sets out the record I build to. No case study on this site comes from banking or from ports. For the port and the companies around it, what transfers is data work. I turned a sync between two operational systems into a data warehouse that keeps history, and that is the base an agent needs before it can answer about anything in motion.

Working across one or two hours

From early November to mid March, Philadelphia is one hour ahead of Monterrey. From mid March to early November it is two hours ahead. Monterrey stays on one hour all year. The code is kept in your repositories, the review is written, and calls are for decisions that need a conversation.

First step and limits

Write with the role or the problem. I study the current process, then reply with a first thing to build, the access it needs and what I would not promise. A senior role or a contract engagement can follow from that. A presence in Philadelphia is not on offer, and neither is advice on regulation.

Common questions

Would I be hiring someone in Philadelphia?

No. I live and work in Monterrey, Mexico, and a company in Philadelphia would work with me remotely. If the job needs a person on your floor or in your hospital, this page is not the answer.

Which hours do Philadelphia and Monterrey share?

Philadelphia is one hour ahead of Monterrey from early November to mid March and two hours ahead from mid March to early November. Nine to five in Philadelphia is eight to four in Monterrey in the first period and seven to three in the second. Most of the working day is common to both.

How would an agent be tested before it is near patient data?

On scripted cases and synthetic users, which is how the evaluation work I publish is built. No case study on this site comes from a hospital. The nearest is a COVID-19 testing platform for a laboratory, operations software that I built in 2021 and 2022, not an AI system. What an agent may reach in production is decided by your privacy and security teams.

Is there published work from a university?

Yes, from higher education. The case study on evaluation gates describes a multi-agent system that serves a higher-education group and handles admissions questions, enrollment, appointments and routing to human teams. It is the closest published work to a university in Philadelphia.

Our data sits in many systems. Where would AI start?

Usually with the data, not with the model. Two write-ups on this site describe that step: an analytics platform that brought fragmented data into one place, and a sync between operational systems that became a data warehouse with history. An agent that reads from a base like that has something reliable to answer from.

By city

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