Engineering
Why, how and with what.
Not a complete timeline — a set of theses on how I think about architecture, tested over twenty years of shipping things to production.
What we're doing right now
We're building a data layer for the various stakeholders in the restaurant market — FMCG, hotels, restaurant groups, public entities — so they can make better decisions based on data. The stack runs on hyperscalers (GCP + Azure), Supabase, Python scripts we build ourselves, and LLM models that are licensed, public, or trained in-house.
I evaluate a lot. I adopt little. What I adopt, stays.
I research technology continuously, across a wide range of stacks — not everything goes live, but a lot does. That constant research is what lets us adopt fast, as pragmatically as possible, whenever a technology actually makes sense for the business.
Multi-tenant all the way down the stack
Laravel, MySQL and Redis running multiple operators on the same system since 2016. Designed that way from the ground up, not bolted on afterward as an isolation layer.
Python since 2018
Scripts, ETL, integrations, automation — since before the LLM wave, which helped prepare the ground for adopting them when they arrived.
And I still write code
In production since 2006 (PHP + a forum). Twenty years later I decide architecture, but I still commit code.