Eduardo CastellanoAI Engineer

AI Engineer · Head of AI Creative Systems & Production

Eduardo Javier
Castellano Rosales

I design agentic AI systems, build AI products, and produce with them: ads, content, documents, sites, music, film, animation and games.

  • AI Engineer
  • Agentic AI systems
  • AI product builder
  • Creative technologist
  • Multidisciplinary producer
NescaféBrand · Product
CelsiusBeverage · Product
TakisFood · Social

I build and lead AI-native creative production systems that turn creative ideas into high-impact content at scale: more speed, more variants and conversion-focused creative.

From my résumé

~30 min

per piece, from idea to finished creative

3 days → 1.5 days → ~30 min (−99%): from manual, to automated, to AI-powered.

Nescafé · Dove · Levi’s

brands served in a creator-economy environment

Plus production for Sears. Commercial, social, UGC, cinematic and more.

33

AI platforms and models tested in real workflows

Compared for quality, consistency, control, speed and production viability.

10 · 388

sectors and services in the catalog I designed

One production system behind every sector on this site.

01

Production in practice.

AI versions produced at Global Media Review Inc. (La Neta / Elevn) to present the AI content service to brands, alongside the deliverables produced with real people.

All work →
NescaféBrand · Product · Commercial creative concept
SearsRetail · Product advertising concept
CelsiusBeverage · Product · Performance-led product creative
LalaFood & Beverage · Product storytelling

02

The system behind the work.

Once the production environment is configured, an AI production agent takes the work from idea to finished creative.

  1. 01

    Creative idea

    Brief and objective: the starting point.

  2. 02

    AI agent

    ChatGPT / Claude: reasoning, planning and agentic intelligence.

  3. 03

    Production environment

    Wireflow and other platforms: capabilities connected to execution.

  4. 04

    Human review

    Validation, judgment and final creative control. Finished, test-ready creative.

AI product builder

Multi-sector AI production platform

I designed and built an AI production platform that serves eleven sectors from one system. The services on this site come from its catalogs.

  • An agent engine in Python and a web app in React 19 + TypeScript, with a landing, a client app and an operator console.
  • Eleven bilingual (EN/ES) sector sites, each with its own service catalog: units, scope rules and guardrails.
  • The work model is part of the product: the client asks, the system proposes, and nothing is delivered until the client approves.

Engineering

What I build with

  • LLM agents
  • MCP, APIs & webhooks
  • Node-based pipelines
  • Cloud execution
  • Generative specification
  • Model evaluation
  • Human-in-the-loop QA
  • Product engineering

Agents configured with instructions, knowledge, state and tools; node-based pipelines on cloud compute; validation and human approval in the path.

Read the architecture →

04

Two ways to work together.

For employers

An AI engineer who also ships the output.

  • Agentic production systems, node-based pipelines and model evaluation.
  • An AI product built end to end: engine, web app and eleven sector offers.
  • Hands-on craft: editing, motion, color, sound and design.

For clients

You set the goal. I deliver the work.

  • I propose the scope; you approve before anything starts.
  • Versions you compare and adjust on the piece itself.
  • Nothing goes out without your approval. AI-made pieces are labeled.

With an account you follow your orders in «My orders». Without one, the form sends it to my email.

Tell me the result you need.

No fixed price: tell me what you need and we’ll agree. I deliver fast and in volume. I reply with a proposal: scope, formats and dates.