Coimbra, Portugal — 2026
Full-Stack Developer & AI Engineer
I build production software that ships.
Different tools, one pattern — each turns raw signals into decisions that make people and businesses measurably more efficient.
FastAPI · Next.js · Databricks SQL · SQLite · Linux VPS
Unifies data from 5+ platforms — each with its own database structure — into one source of truth across 1,848+ accounts, then turns it into KPIs that tell account managers what to do next and why, not just what the numbers say. Built-in tools let non-technical staff run technical operations they couldn't before it existed. In production daily.
Account data was scattered across 5+ platforms, each with a different database structure — no single source of truth, just raw numbers nobody had time to interpret, and technical tasks that always needed an engineer.
Fetch and reconcile all of it into one model, compute the KPIs that matter, and turn them into prescriptions — 'do X for account A because of B' — instead of just charts. One example: reading sensor patterns to tell a failing sensor from a dead device, a distinction that didn't exist before.
Account managers act on guidance instead of digging through five dashboards, and non-technical staff perform technical tasks themselves. The platform decides what deserves attention; people just execute.
Chrome Extension · TypeScript · REST APIs
A Chrome extension built for the Anova Customer Success team that wraps complex technical operations into simple, one-click actions — so non-technical agents can resolve issues themselves instead of escalating to engineering or another department.
CS agents kept hitting technical tasks beyond their tools — checks, lookups, and fixes that meant stopping to escalate to engineering or another team, leaving the customer waiting and burning cross-team time.
Wrap those technical operations into one-click actions inside the browser agents already work in — hide the complexity, surface only the decision they need to make.
Non-technical agents resolve far more on their own, escalations drop, and customers get answers in the same conversation instead of waiting on another department.
Next.js · Express.js · PostgreSQL · TypeScript
Stops small Portuguese beauty and wellness businesses from losing bookings to phone-tag and no-shows — clients book themselves 24/7 while the system runs the calendar owners used to manage by hand. Live with real paying customers, in PT/EN/ES.
Small Portuguese beauty and wellness businesses lose bookings to phone tag and no-shows, and owners burn hours a week running the calendar by hand.
Treat the calendar as data — a self-serve booking layer that turns every appointment, reminder, and open slot into something the system manages and the owner can read at a glance.
Owners reclaim hours every week, clients book 24/7, and no-shows fall — the business runs leaner without hiring.
React · TypeScript · Chrome Extension · Node.js · PostgreSQL
A React browser extension that lets affiliate marketers turn any Amazon, Walmart, or Target product into a ready-to-ship post in seconds — and schedule them ahead. A PostgreSQL data layer powers analytics that reveal how products and the affiliate's own niche are performing. The React UI lives right in the browser, where the work happens.
Affiliate marketers lose hours hand-building product posts one at a time, with little visibility into which products or niches are actually worth promoting.
Make posting a one-click pipeline — pull a product, generate the post, schedule it — and log everything into a PostgreSQL dataset whose analytics reveal the wider market and the affiliate's own niche in depth.
Affiliates ship far more posts in far less time, schedule ahead, and decide what to promote from data instead of gut feel.
I don't have a degree yet — I have deployed systems that real companies depend on. Over the past year I built a full B2B SaaS analytics platform from scratch (used daily at Anova), a booking SaaS with paying customers, and a browser extension for e-commerce price intelligence.
I work across the full stack — Python and FastAPI on the backend, Next.js and TypeScript on the frontend, Databricks and PostgreSQL for data, Linux and systemd for deployment. When there's a problem, I build the solution. That's the only way I know how to work.
Quick facts