Santiago de Compostela, Galicia, Spain
Most AI systems look brilliant in a demo. But when they hit the messy reality of multi-language, high-scale enterprise production workflows? They break. Fast. 💥 That's where I come in. I don’t just build AI. I build reliable AI that global enterprises can trust. As an AI Engineer and the Executive Director of Interkultural, I bridge the gap between human language complexity and production-ready systems. My secret weapon? A background that uniquely combines journalism and translation with rigorous data science engineering and a high level of cultural intelligence. In the world of LLMs, linguistic context isn't an abstract concept—it’s a strict engineering constraint that must be modeled, evaluated, and controlled. 🌐 🛠️ The Architecture of Reliability I design the orchestration and evaluation layers that turn unpredictable LLMs into dependable production systems. Here is where I focus: 🤖 Agentic Workflows & Orchestration: Designing multi-step reasoning frameworks using LangChain and LangGraph. I build autonomous agents capable of planning, executing, and self-correcting across complex enterprise tasks—moving way beyond basic conversational interfaces. 🔍 Advanced RAG Systems: Structuring Retrieval-Augmented Generation pipelines with surgical precision. By optimizing chunking strategies, embedding selection, hybrid search, and reranking, I engineer architectures that eliminate hallucinations and maximize semantic grounding. 🌍 Multilingual NLP Architecture: Solving silent failure modes in non-English systems (including English, Spanish, Portuguese, and Galician). I design systems that eliminate tokenization gaps, semantic drift, and the loss of cultural nuance so reasoning remains flawless across borders. ⚙️ LLMOps & Production Stability: Implementing rigorous evaluation frameworks (RAGAS, TruLens), prompt versioning, and enterprise vector data systems to ensure AI performance is scalable, stable, and entirely measurable. 💻 Technical Stack - Languages & Core: Python, SQL, NLP - Frameworks: LangChain, LangGraph, LlamaIndex, Hugging Face, PyTorch - Data & Infrastructure: Vector databases (Pinecone, Milvus, Weaviate), API-first design, cloud-native deployments - Focus: Architectural rigor, automated evaluation, and production stability 🤝 Let's Connect I am looking for my next challenge as a Lead AI Engineer or AI Architect within enterprise or high-growth environments. If you want to build AI systems that are not just intelligent—but bulletproof, measurable, and ready for global scale... 👉 Slide into my DMs and let’s talk! 🚀
As Executive Director, I don’t just sit in a boardroom. I’m in the trenches, writing code and building the core infrastructure that makes enterprise AI production-ready. 🛠️ At Interkultural, we’re focused on turning volatile LLMs into high-performance enterprise systems. I lead technical execution, engineering the agent systems and data configurations required to deploy reliable AI across complex, high-variability linguistic environments with a high level of cultural intelligence. Here’s what my hands-on architectural deployment looks like: 🤖 Agentic Infrastructures: Designing autonomous reasoning systems using LangChain and LangGraph. I implement state management, persistent memory, and tool-calling interfaces that let agents plan, execute, and self-correct across multi-step enterprise workflows. 🔍 Advanced RAG: Developing high-precision pipelines. I specialize in chunking, embedding selection (dense/sparse), hybrid retrieval, and cross-encoder re-ranking inside Vector Databases (Pinecone, Milvus) to eliminate hallucinations. 🌐 Multilingual NLP: Eliminating failure modes in non-English systems. I build architectures that solve tokenization gaps and semantic drift, ensuring flawless reasoning consistency across English, Spanish, Portuguese, and Galician. 📊 LLMOps & Validation: Implementing evaluation frameworks (RAGAS, TruLens) and prompt versioning. I focus entirely on technical stability, ensuring deployed solutions are measurable, scalable, and optimized for production. 💡 The Engineering Edge: My background in journalism and translation teaches me to treat text data like structured code. I treat linguistic ambiguity not as an abstract concept, but as a functional engineering constraint. 💻 Core Stack: LangChain, LangGraph, LlamaIndex, PyTorch, Python, SQL, NLP, Pinecone, Milvus, Weaviate. 🚀 Let's Connect: Scaling reliable AI or looking for a Lead AI Engineer/Architect? Let's talk. Drop a DM! 🤝
Long before LLMs were a concept, I spent a quarter-century studying, debugging, and structuring human language data. The reality of AI is simple: you can't build a reliable enterprise AI pipeline if you don't deeply understand the high-variability data going into it. 💥 Throughout this 25-year era, I operated at the intersection of language structure, global localization, and cross-border business logic. This gave me a deep understanding of international markets, localization challenges, and global enterprise communication. This is how that execution shaped my architectural foundation: 📰 Information Architecture & Journalism Focused on data verification, syntax precision, and structural mapping. This built my systematic approach to how data must be validated. 🗣️ International Brand Translation Directed high-level communication strategy and cross-cultural adaptation across competitive global markets. I didn't just translate words—I engineered how core brand messaging and intent map safely across cultural boundaries without losing logical integrity. 🌐 Intercultural Communication Consulting Designed communication frameworks and modeled abstract context as a strict engineering constraint for global businesses. I advised multinational entities on navigating complex institutional, linguistic, and cross-border communication barriers to scale safely. 🧠 Linguistic Foundations for AI This multi-decade era was the ultimate R&D prerequisite for AI engineering. By mastering the deep structural rules and semantic variations across English, Spanish, Portuguese, and Galician, I built the exact mental models required today for advanced prompt architecture. 🚀 This 25-year mastery of language data is the exact foundation that powers my current architectural work in multilingual Generative AI. I know how global markets communicate, and I know how to write the rules to automate it. 💬 Let's connect and talk strategy. Drop a DM! 🤝
Cultural stories require precision and style. I collaborated with journalist Aser Álvarez to execute a major cultural feature with Xosé Luís Méndez Ferrín—one of contemporary literature’s most singular intellectual voices. The interview became the main cover piece for the cultural supplement "Sermos Galiza" in "Nós Diario." In practice: 📰 1. Editorial Contribution This wasn't just a basic interview. It was high-stakes information synthesis. We managed deep biographical research, joint narrative planning, and rapid source verification to translate a multi-hour dialogue into a cohesive, heavy-hitting feature. 🗺️ 2. Biographical Storytelling We structured the entire narrative around the concept of the raia (the borderland). Instead of a mere geographical space, we modeled it as a rich historical territory—mapping complex, abstract concepts of identity into high-impact journalism. 🗣️ 3. Tonal Voice Calibration Preserving stylistic cadence and tonal accuracy under tight constraints is everything. Much like fine-tuning an AI model's temperature and voice today, my job was to ensure the text fully retained the profound intellectual density and unique expressive cadence of Ferrín's voice. 🔮 4. Linguistic Fluency Conducted and published entirely in Galician. This required native fluency and deep cultural awareness, treating the language not just as a medium, but as a living space of meaning and continuity within the national literary conversation. 🧠 The Linguistic Takeaway: To build great tech, you must first master complex human concepts. My ability to digest highly advanced, abstract ideas and structure them cleanly is the exact skill set I use today to map, control, and optimize complex enterprise language data.
Building a brand-new, self-sustaining global organization from scratch is tough. Doing it with a 100% volunteer team? That requires serious leadership skills. 💥 When we set out to launch a brand-new branch of Toastmasters International, we didn't just want another online meeting room. We wanted a high-performing ecosystem that could scale without losing its edge. As a founding team member, my job was to build an infrastructure that kept a distributed volunteer workforce highly motivated and perfectly aligned. The operational summary: 📐 1. Governance & Compliance Architected the operational frameworks, governance protocols, and internal standards from day one. We designed clear leadership models and ironclad guidelines. This structural discipline is exactly what helped our decentralized team clear every regulatory hurdle and secure our official international certification. 🔄 2. High-Performance Loops Deployed continuous optimization loops and rapid peer evaluation mechanisms. Because volunteers trade their valuable time for personal growth, we built a safe training environment. By embedding fast, constructive feedback loops directly into our rhythm, we turned a distributed group into an agile, highly accountable network. 🌐 3. Volunteer Mobilization & Growth Led strategic growth planning, member retention engineering, and fully remote coordination across international time zones. Leading people who don't work for a paycheck takes deep execution discipline and sharp empathy. We built communication processes that kept everyone inspired and moving forward together under evolving conditions. 🧠 The Systems Leadership Takeaway: Elite communication and structural design are two sides of the same coin. Building this initiative from absolute zero with a volunteer team proved a massive truth: when you set up the rules of engagement correctly, you create a space where global teams can execute, scale, and thrive smoothly.
If you can't capture attention across time zones, your organization doesn't exist. Great PR isn't about making noise. It's about building a message that drives real action. 💥 As the VP of Public Relations for a global, fully online leadership community, I didn't just pass along information. I built the narrative ecosystem that scaled our visibility across borders while strictly optimizing our resources. Putting it into action: 📢 1. Strategic Communication Architected global brand positioning and end-to-end multi-channel communication strategies. By carefully optimizing our resources within a volunteer framework, we built an ironclad digital presence that stabilized our identity during a critical scaling phase. 👁️ 2. Visibility Engineering Executed high-impact community growth strategies to scale our digital visibility. We systematically turned casual observers into active, committed members by shifting the narrative from basic public speaking to high-value personal development and confidence. 🌍 3. Global Tone Alignement Managed digital PR channels and cross-border coordination across multiple global time zones. Operating in a fully remote, multicultural ecosystem required absolute sensitivity to tone, ensuring our messaging remained accessible, professional, and resonant across distinct demographics. 🎯 4. Lifecycle Engagement Oversaw brand voice optimization by creating strict, consistent messaging blueprints across all digital touchpoints. I treated language as a mechanism for community alignment, reinforcing a culture of discipline, feedback, and collective growth. 🧠 The Influence Takeaway: How you shape meaning dictates how you scale. Mastering global PR proved that when you control the narrative and optimize your resources, you build the absolute trust needed to drive massive participation in a digital-first world.
Translating a global publication coordinated by marketing expert Seth Godin isn't just about changing words. It's about high-velocity semantic data orchestration across a decentralized global network. 💥 "El almanaque del carbono" was a comprehensive collective intervention in how climate data is understood worldwide. My role was to bridge the gap between high-density technical frameworks and international market psychology alongside elite global players. The strategy: 📊 1. Technical Localization Managed highly dense, multi-disciplinary technical knowledge spanning science, economics, and ecology. I ensured absolute semantic precision, transforming complex environmental datasets into accurate, localized information assets without distortion. 🌐 2. Distributed Operations Executed high-velocity agile project coordination within a decentralized global network. Operating as a critical node in an international, distributed editorial process required strict methodological consistency, rapid alignment, and continuous quality loops. 🎯 3. Semantic Integrity Eliminated semantic drift and maintained total structural integrity across distinct linguistic systems. I carefully calibrated terminology, layout, and tone across multiple contributors to preserve density of meaning without compromising clarity. 🧠 4. Cognitive Accessibility Aligned the technical material with high-level global marketing logic, user experience (UX), and behavioral psychology. We didn't just translate data—we optimized it to drive real user engagement and collective responsibility. 🌍 The Global Strategy Takeaway: Working with the best in the business taught me a fundamental truth: translating data without losing intent is exactly how you scale complex systems safely across global markets.