I’m a senior backend developer who’s been building internet software since 1997 — started on a Z-80 clone at age 10, never stopped.
I’m based in Brazil and I fell in love with open source almost two decades ago. Drupal has been my home since — I maintain several contributed modules, fix security vulnerabilities, and mentor new contributors. I also spend a lot of time in Tmux and Neovim these days.
For a full resume with my work history, certifications, and project portfolio, check out the resume page.
Local LLMs & Open Source AI
Lately I’ve been spending more time than I should pushing models to their limits on consumer hardware. I’m fascinated by what open-source AI can do when you control the whole stack.
My setup: an AMD RX 7900 XT (20 GB VRAM), an Intel i7-14700K, and 64 GB of RAM. I run Qwen 3.6 27B via llama.cpp with turbo quantization, getting 29 tokens/sec decode speed at 132K context. Everything local, everything open source, zero cloud dependencies.
I believe local LLMs are production-ready for real development work. This blog is proof: it was built entirely with a local Qwen model through Pi Coding Agent, in a single session, without sending a single byte to the cloud.
Outside of Code
When I’m not behind a keyboard, I’m probably on my motorcycle or picking up a guitar.
Contact
Find me on GitHub, Drupal.org, or send an email.