Built with AI
What practical AI fluency looks like.
Practical AI fluency is the ability to turn a real problem into a working system—not merely to write prompts.
I define the problem, expected behaviour, constraints, information sources, and success criteria, then use AI to implement, test, and iterate. The projects below started as recurring friction in my own work and became tools that run.
I am not presenting myself as a software engineer. My strength is knowing what should be built, how the parts should work together, where AI is useful, and how to keep iterating until the result solves the original problem.
Job Hunt Automation
My own job search, turned into a pipeline. It pulls roles from several sources, judges fit with a mix of local and cloud models, scores my CV against each posting, and sends only the survivors to a private dashboard. Then I rebuilt a clean, public version anyone can run, fully offline: JobScout, open source on GitHub. It's the one that actually found me roles.
Metro Agora
Real-time arrival countdowns for the Lisbon Metro, wrapped around a custom proxy that handles the official API's quirks. It's the one build here that's public: anyone can use it at metroagora.com, in the browser or installed to a phone's home screen.
Fogos Firewatch
A wildfire alerter built for Portuguese summers. Most fire feeds buzz you about everything burning in the country. This one weighs wind direction and distance, and only speaks up when a fire is genuinely heading your way.
These three are the clearest examples. I've built others the same way: earthquake alerts, finance dashboards, a PR-triage bot, each one a real problem I solved in an afternoon with AI doing the heavy lifting.