Gabriele Lucci

Software developer

I’m a software developer with professional experience in Python and JavaScript/Node.js, particularly on the backend side of web development, with a proven track record on delivering cloud-native solutions from the drawing board to production.

I’m a detail-oriented perfectionist, but also versatile and constantly looking for new challenges. I enjoy working with Django (who doesn’t? ), but I’m always open to learning new technologies. Additionally, I’m quite proficient in administering Linux systems, which I’ve been using in some form or another since I was a kid.

In my spare time, I especially enjoy engaging in various music-related activities, such as playing bass guitar with my friends.


Work Experience

Technician / Full stack developer

May 2021 - Dec 2023
Eurac Research (Institute of Biomedicine)

As part of the IT team at the Institute for Biomedicine of Eurac Research, I worked primarily on the CHRIS-2D project, which involved the creation of a “dynamic data portal”, a web app enabling researchers, even from outside institutions, to peruse a set of the data gathered by Eurac’s IfB studies.

Due to the small size of our development team, we decided to build the portal as a full-stack Django application, that included an HTTP API to be queried by complementary internal tools. I played a significant role in deciding the project’s architecture and implementation; it’s been a real benchmark of my skills that was ultimately a success, and I’m quite proud of what was achieved.

Backend developer

Jul 2017 - Jan 2021
Fondazione Openpolis

I worked for the Openpolis Foundation, supporting their editorial team by developing multiple web applications and toolkits using Python and Django.

Those are the most relevant projects I was involed in:

  • Mappe del potere”: essentially a database of all things related to Italian politicians, modeled following the Popolo open specification. It essentially consists of a core implemented as an open-source Django app, a complex REST API that exposes the underlying data, and a custom-built task manager to handle and scheduling background tasks. The data was gathered from numerous different sources (HTTP APIs, SPARQL, web scraping) by ETL pipelines.
  • Povertà educativa”: a project that shares a lot of components and concepts with the previously listed one, but with a particular focus on the socio-economic conditions of minors in Italy.
  • A schede scoperte”: an application implemented in Django and React, with the purpose of promoting transparency during the 2018 Italian general elections.

This was a pivotal role for my career, as I was lucky enough to have an expert mentor who taught me a lot of extremely valuable skills that I consider fundamental to my profile.

Full stack developer

Sep 2016 - Jul 2017
Anansi Team
I was the lead developer of “APA Cloud”, a suite of server side applications that would gather, store, and visualize air quality sensor data. Mainly developed using Node.js (LoopBack 3 and Sails.js) for the backend, and Angular v1 for the frontend.

Projects

My own self-hosted "cloud"

a.k.a. "LucciCloud"

I’ve been maintaining an array of self-hosted services for years.

Some of the services I run (there are more, but these are the most relevant):

  • a bunch of Fediverse nodes (Mastodon, GoToSocial)
  • a Nextcloud server, which I use daily to back up my photos, store documents and sync contacts and calendars between my personal devices.
  • Grocy, an ERP for managing groceries and alike.
  • a Minecraft game server because… why not?
  • this very website (made with Hugo)!

Everything is deployed and orchestrated using containers on rented Linux hosts. Services that expose HTTP functionalities are proxied using Caddy, which automatically handles certificates and also acts as load balancer. Additionally, pretty much everything is “instrumented” using Prometheus and monitored using Grafana.

I’m very proud of this project, it taught me plenty of skills that complement my developer profile quite well, and I consider it a great way to keep up with the latest trends in the field.

Private "AI lab"

Experimenting with local LLMs

I’ve been experimenting with local LLMs for a while now, using llama.cpp and various complementary tools to run models locally on my own hardware. I have a few different models running, that I employ for different tasks, such as Llama for conversational tasks, Deepseek Coder for Copilot-like code generation and completion, Granite for text embedding of documents, and others.

To make the usage of these models easier, I employed simple frontend solutions such as Open WebUI and ComfyUI, which allow me to query the models using a web interface.

This is a fun project that helps me to keep up with the latest developments in the field of AI. It also provided me with a basic understanding of what it takes to deploy local models, a skill that could be useful in contexts where data privacy is a concern and the use of cloud services isn’t desirable.