I'm a linguist and I optimize agentic AI to execute with understanding - not improvise.

I work building and auditing the context and rules that govern its behavior: so it produces what the automation is expected to produce, within the real limits of what agentic AI can do.

That's done through language, not code. I work on what feeds the agent: the context it receives, the sources it relies on, the policies that constrain it, the system prompts, workflow definitions, and tool definitions that structure its behavior, and the guardrails that keep it in place.

What's behind

Fifteen years working the same problem in different forms: how machines process language. First search engines, now agents. SEO, AEO, schema markup, semantic architecture - the same question applied at different scales: how do algorithms interpret meaning, and what happens when they don't. Schema markup is structured semantics for crawlers. AEO is structured semantics for AI search. Agentic AI is the same problem at a different scale of autonomy. Agentic AI is where the career comes full circle: back to the original question, how semantics impact meaning and conversations, only now the interlocutor is an agent. The same knowledge base that shaped how these models were built is what makes it possible to diagnose why they fail.


Degree in Linguistics (extraordinary achievement award). Fulbright Scholarship, Wheaton College (Massachusetts) - research in how language impacts meaning and conversations when mediated by machines. That academic foundation is what makes this work possible.

How I trained in AI

Ongoing formal development alongside practice:

Stanford Online credential - AI-Driven Leadership: Strategies for the Future, completed by Elena Benito Ruiz, October 2025
Stanford Online · Stanford Engineering
AI-Driven Leadership: Strategies for the Future
Certificate of Achievement · October 2025
Verify credential →
Certificate of Completion - AI Governance Training by Luiza Jarovsky PhD, issued to Maria Elena Benito Ruiz, October 2025
AI, Tech & Privacy Academy · Luiza Jarovsky, PhD
AI Governance Training
Certificate of Completion · October 2025

I have studied in depth the training and learning of machine learning algorithms with Andrej Karpathy, papers and specialized journals.

I also write The Semantic Agent, a newsletter on semantic architecture and agentic AI.

My position on AI

Designing how AI agents behave requires a clear position on what the technology is for. I work with AI because I understand the mechanism - and with that understanding comes a clear view of what it should and should not replace.

Read my full position →