Why and how I built webtelemetry.dev: a live survey of new websites, and my way of learning how to process large amounts of data with Kafka and PostgreSQL. Plus what it found about AI crawlers and llms.txt.
I was tired of checking GA4, Search Console, and Bing in separate tabs for every site. Here is why I built one private control center for the whole portfolio.
How I used Claude Code to analyze the Grafana repository and generate a complete onboarding page with architecture diagrams, role-based views, and interactive navigation.
How Utility Forge uses Product Owner, Software Engineer, and QA agents coordinating through GitHub Issues and Actions to ship developer tools end-to-end, every day.
A practical look at my MCP server that combines Google Search Console and Google Analytics, letting you ask SEO and traffic questions in plain English.
How I built a web app where four AI agents with distinct personalities debate your decisions in real-time — covering streaming, prompting, and verdict parsing.
How to configure Claude Code’s agentic coding workflow to run on free local Ollama models instead of paid API calls — setup, model picks, and tradeoffs.
How I built PocketLLM — an Android app that runs Google’s Gemma 3n model fully offline using MediaPipe Tasks GenAI, with architecture, code, and performance notes.
Why local AI matters and how to run Ollama + Hugging Face models on your laptop — covering installation, model selection, quantization, and API integration.
An intro to the Model Context Protocol: how it lets AI tools like Cursor query a Postgres database in natural language, with a full integration walkthrough.