New York is the financial center of the country, and it is also known for media and advertising, law and professional services, healthcare, and fashion and retail. I work with companies there from Monterrey, Mexico. This page is written for someone who has to decide quickly whether my work fits their industry, so each section names the need, the service and the evidence, and says so when the evidence comes from somewhere else.
Finance
In finance the model is rarely the hard part. Approval is. A bank or a fund wants to know what an assistant can reach, who allowed it, and how anyone would notice if its answers got worse. No case study on this site comes from banking. What transfers is the control itself, shown in a gateway on Cloud Run where many tools sit behind one governed entry, with authentication around each call and evidence after it. The service is enterprise AI, and there is a guide on enterprise agent architecture for the wider design.
Media and advertising
A publisher holds an archive, and an agency lives on new business. For the archive the need is retrieval that reaches the right source and can be checked, which is the work of RAG and data platforms. For new business the published work is outbound engineering: a domain that earns the inbox before it sends, a prospecting engine in which every claim about a company carries where it came from, and a funnel that receives the answer. The services are email automation with AI and AI sales development.
Law and professional services
A law firm cannot send a client a sentence that nobody can trace. No case study on this site comes from a law firm. What transfers is a rule from my outbound and evaluation work: a claim without a source does not go out. In a law firm that would mean answers drawn from a controlled set of documents, with the passage shown, and tests that fail when the agent states something the documents do not say. RAG and data platforms and LLM evaluation are the services.
Healthcare
Hospitals ask for what banks ask for, under different regulators: narrow access, a record of every action, and proof that a new version is no worse. The published work that transfers is a set of evaluation gates for a multi-agent system that serves a higher-education group with 700,000+ students. It is education, not medicine. The method is what carries over, and LLM evaluation is the service.
Fashion and retail
A retailer answers the public in volume and keeps product data in more systems than it would like. No case study on this site comes from fashion or retail. What transfers is data work, an analytics platform and a data warehouse fed from operational systems, and agents built on top of records that are already consistent.
One or two hours behind
Monterrey keeps one hour all year and New York does not. New York is one hour ahead of Monterrey from early November to mid March, and two hours ahead from mid March to early November. Calls fit best from late morning in New York. The rest moves through pull requests and written notes.
Before you write
I work under my own name, in a senior role or on contract. Send the problem by email, or choose a time for a call on the contact page. What you receive first is a written reading of your process. What you will not receive is an engineer in New York.