Aquí es donde comparto ideas, conocimientos y lecciones que he aprendido en el camino. Ya sea algo que haya aprendido a través de la experiencia, la investigación o la curiosidad, espero que encuentres valor en lo que publico aquí.
Beyond the Hype: Why Foundations Outperform "Vibe Coding."
Is your AI strategy built on solid engineering or just lucky prompts? Discover why understanding the math behind the models is the only way to build production-ready AI that scales without the "noise" of the latest trends.
Multi-Agent Systems: Architectural Lessons from the Great Tomb of Nazarick
Is your AI being too "loyal" to be useful? Explore why the "10,000-year plan" in Overlord is actually a masterclass in AI Sycophancy and how to manage the "Cold Logic" of autonomous Multi-Agent Systems.
Why Your "Perfect" Spec Files Are Making Your AI Agents Worse
Is your AI agent hallucinating despite your detailed specs? More context isn't always better. Learn why keeping your spec files lean and delegating to specialized agents is the only way to scale specs-driven development without the "context bloat" trap.
From Synthesizer to Supervisor: The Sovereign Architect in the Age of Agentic Workflows
The age of the "coder" is over; the age of the Architect has arrived. As AI moves from basic autocomplete to autonomous agentic workflows, the role of the Senior Engineer is shifting from writing logic to auditing intent. Are you prepared to manage a team of synthetic contributors, or will your codebase succumb to architectural drift?
Stop Prompting Pixels, Start Coding Images
Discover why treating images as code, not pixels, is the future of design. Learn how SVG and vector-based workflows provide superior control, scalability, and performance compared to AI-generated raster images.
Insights from SALA: The LatAm AI Summit
A recap of the SALA LatAm AI Summit, featuring insights on computer vision, sustainability, reasoning, and the future of LLMs.
Don't Learn Claude Code, Learn This Instead!
Instead of tightly coupling your skills to a specific AI model like Claude, focus on mastering foundational deep learning concepts, robust evaluation metrics, and flexible system design. This model-agnostic approach ensures you can adapt quickly as new tools emerge, keeping your expertise relevant and your workflows resilient in a rapidly evolving AI landscape.
Delivering What Matters: The New Standard for AI-Era Web Experiences
As AI rapidly transforms how content is created and consumed, traditional content management systems (CMS) must evolve to keep pace. This article explores the challenges of information overload, the shift in user behavior toward AI-driven content synthesis, and the limitations of conventional CMS interfaces.
Manage Your Flat File Based Website Content with Decap CMS (Formerly Netlify CMS)
Ever wish managing your content site was as smooth as your git commits? If you’re tired of bloated, database-heavy CMS platforms or manually editing Markdown files every time you need to publish a post, you’re not alone. Enter **Decap CMS**—a powerful, git-based, headless CMS that’s built for speed, simplicity, and modern web workflows. If you're using static site generators (SSGs) like Hugo, Jekyll, or Eleventy, or building with server-side rendering (SSR) frameworks like Next.js, this is the content management solution you’ve been looking for.
Deploy your statically generated site to a Cloudflare Worker
If you're building a static site (or even a SSR one) with tools like Astro, Hugo, or Next.js, you can supercharge its performance by deploying it to the edge with Cloudflare Workers. This post walks you through the entire process, from generating your site and setting up a Worker project with wrangler, to configuring your site and going live. You'll also learn how to set up a custom domain. It's a step-by-step guide to getting your site fast, secure, and live on the edge in minutes!
Host Your Own GitLab Repository Server on Linux (Debian 12)
When working with development tools, having a place to store and manage your code is essential. While many cloud-based services are available, a self-hosted solution offers greater control over versioning and CI/CD pipelines. You can achieve this by setting up your own GitLab server, providing full ownership and customization of your repositories and workflows.
