The Blog
The AI Engineering Blog
Deep dives, tutorials, and field notes from the lab - building agentic AI, local-first LLM infrastructure, and self-hosted systems in production.
Introducing Gekro News: An AI Briefing That Curates Itself Around Me
A public daily AI briefing that reads a profile of my interests distilled from my own knowledge base, picks the day's real signal, cites its sources, and publishes itself every morning.
The Token Economics of Local AI
Why every team running sustained AI workloads - from a solo home lab to a 50,000-user enterprise - pays itself back faster by owning the inference layer than by renting it.
Sonic Phoenix: Bringing 7,246 Forgotten Songs Back from the Dead
How my child's request to hear my childhood music turned into a seven-phase pipeline that fingerprinted, sorted, enriched, and synced a 30GB library to Spotify - then handed it to an AI skill for on-demand playlist curation.
AI Codes Like a Genius. Architects Like a Goldfish.
Why zero-shot AI POCs fall apart at scale, and how constrained architectural templates are the only thing standing between a clean codebase and a production time bomb.
Financially Air-Gapping A Raspberry Pi AI Swarm
How I turned a Pi 5 into a localized orchestrator using an MCP router and isolated OpenClaw worker networks to slash API costs.
Printing the Foundation: 3D Construction as Green Infrastructure
How additive manufacturing is disrupting material science and the traditional role of the contractor.
Terminal Velocity: The CLI as your AI Abstraction Layer
Why GUIs are a bottleneck for AI engineering and how to build a high-speed command line workflow using WSL2 and Zsh.
Hello, Ollama: Local Inference is Your Architectural Insurance
Running LLMs on a Raspberry Pi isn't just a hobby; it's a fallback strategy for system resilience.
API Sovereignty: Building for the 2 AM Failure
Why generic API wrappers are a liability and how to build a resilient, multi-provider fallback chain.
The Mac Mini M4: The Un-official Local LLM King
Why unified memory architecture is the only way to run 70B parameter models without a data-center budget.
The Linux Edge: Why AI Breathes Better in the Kernel
Why I stopped fighting Windows registry errors and moved my entire AI engineering lab to WSL2 and Ubuntu Server.
The Raspberry Pi Master: Your 24/7 Lab Assistant
Why my Pi isn't just for hobbyist projects, but serves as the always-on utility and IoT bridge for the Gekro Lab.
The Architecture of a Lab: Splitting Brain from Body
Why monolithic web frameworks fail for AI engineering and how I built a distributed, agent-first architecture for Gekro.
Zero to Agent: My First Autonomous Loop
The moment I stopped writing scripts and started building systems that sense, think, and act without my permission.