Production-grade software, engineered end to end — architecture, code, and delivery for new products and the systems already running.
I handle software at every layer — from modeling the data and background services to polishing the user interface, applying the same engineering standards to every step.
Production software shipped — from real-time IoT platforms to ERP systems, SaaS products, and client sites — across front end, back end, and the infrastructure between, adapting to whatever each project demanded.
work/
labs/
Vrandalismo · Creative Studio · Freelance
Built the marketing site for Vrandalismo, a creative studio, with Astro for fast static delivery and Tailwind for a tightly controlled design system. Wired up Sanity CMS so the team can edit services and case studies on their own, no developer in the loop.
DEHN · IoT Platform · Current
Lead frontend and backend contributor on real-time IoT dashboards and a windmill lightning-detection system, built with Vue 3 and NestJS.
DEHN · Hardware Emulation
Built a virtual device with FastAPI and Vue to emulate proprietary hardware, letting the team verify end-to-end connectivity with ThingsBoard without the physical device.
Essentialist · Luxury Travel
Worked within the team on the backoffice for travel-itinerary management on Angular (14–18) and contributed to the premium customer-facing app for the luxury travel sector. Took part in architectural decisions for scalability and helped run error monitoring with Sentry — triage, fix prioritization, and resolution tracking across releases.
Evalua · Insurance SaaS
Contributed to building and maintaining an Angular SaaS platform for corporate insurance management, helping establish the CSS architecture alongside the team.
Evalua · Insurance PWA
Shipped features for a client-side Progressive Web App that let policyholders manage their insurance and file claims directly.
Bosonit · Internal ERP-CRM
Refactored and developed the company's internal platform using Angular. Customized Angular Material to build a proprietary component library, unifying the UI and speeding up future feature delivery.
Ginzo Technologies · Custom ERP
Contributed to a custom ERP for a client company, building frontend components and responsive interfaces with Angular and the Fuse template alongside the development team.
Script · 2026
A script that pulls the audio track out of videos and transcribes it through OpenAI's API. Built in JavaScript as a focused tool for turning raw footage into searchable text.
Ecosystem · 2026
An ecosystem for signing and verifying W3C Verifiable Credentials 2.0, made of a backend service and a companion mobile app that work together to issue and validate credentials.
SaaS · 2026
A SaaS for couples planning their wedding: generate digital invitations, track guest attendance, and keep the budget under control, all from one place.
Approach
I choose technology based on project requirements, not personal hype. Languages and frameworks are tools: I adapt to whatever is needed to be productive from day one.
Tests and CI aren't an afterthought for the end of the project; they are my safety net to iterate fast and without fear. If breaking things in production has a real cost, testing is mandatory.
I optimize code so the team inheriting it can understand it seamlessly, not to feel like the smartest developer in the room today. Clear contracts, clean architecture, and straightforward decisions.
Working with AI
AI changes how fast we write code, not the responsibility to do it right. The real value isn't in knowing how to write a prompt — it's in the software architecture that allows a model to interact with a complex codebase safely and predictably.
I package business logic and workflows into reusable skills. This ensures AI agents follow standardized processes instead of improvising on every run.
I design the runtime environment and boundaries around the model — including custom tools, validation layers, and sandboxing — to guarantee predictable, production-ready outputs, moving past toy demos.
I write code and documentation optimized for LLM comprehension: strict typing, explicit contracts, and semantic architecture. Conveniently, what makes life easier for the model also makes it easier for your developers.