Senior Software Engineer · AI-Powered
I build AI-powered software: 20 years of backend engineering, now applied to systems that combine solid architecture with LLMs, computer vision and modern ML. I use AI coding tools (Claude Code, Cursor) as part of my daily development workflow.
20+ YEARS
BUILDING SOFTWARE
AI-NATIVE
DEVELOPER
About
I’m Lluís, a Senior Software Engineer with two decades of experience building backend systems and the infrastructure to run them reliably. Today, AI is central to how I work: I use Claude Code and Cursor daily, build applications powered by LLMs and computer vision, and run a self-hosted AI agent that automates parts of my own workflow.
My technical grounding in AI comes from two UPC postgraduate programs (750 hours) in Deep Learning and AI applied to Logistics — completed while shipping personal projects in production.
I believe the best way to understand a technology is to run it yourself. My home server hosts Docker containers, a CI/CD pipeline on GitHub Actions, n8n workflows and a personal AI assistant. I'm curious, hands-on, and I bring things from idea to production.
Skills & Resume
Backend Engineering
Cloud & DevOps
AI Developer Tools
Machine Learning
Databases
2018 — 2024
Software Engineer
2013 — 2018
Software Engineer
2010 — 2013
Software Engineer
2008 — 2010
Embedded / Systems Engineer
2006 — 2008
Software Developer
Projects
A collection of recent AI-driven projects exploring Deep Learning, Generative Models and applied engineering.
iOS/Android app that generates personalised illustrated bedtime stories in Catalan for children using generative AI. A 7-step wizard configures the protagonist, world and virtue; stories are generated asynchronously beat by beat with client-side polling.
Full-stack web app comparing 2 Hugging Face models (RMBG-1.4 and RMBG-2.0) for background removal with side-by-side results and transparent background download. FastAPI backend, Vue 3 + TypeScript + Tailwind CSS frontend, Dockerised with volume-based model caching and health checks.
ML system predicting 3 targets per shipment (load, weight and profit) using XGBoost, with knapsack-type optimisation for truck assignment. Data infrastructure on PostgreSQL with PostGIS (geospatial) and pgRouting (routing). Full-stack with interactive visualisation and live demo.
Interactive fitness platform combining real-time hardware sensors, web software and AI-generated games. An OpenAI assistant generates and auto-repairs game definitions on the fly. Device communication via MQTT/Mosquitto.
Computer vision pipeline for analysing soccer match footage. Detects and tracks all players per frame, clusters them into two teams via HSV color histograms and KMeans, and outputs an annotated video with per-team bounding boxes.
Self-hosted RAG chatbot answering questions about my professional background. Built with Flowise (LangChain), vector search and prompt engineering over CV and portfolio data. Available live in the Ask Me section of this site.
Self-hosted n8n workflow using an AI agent to triage incoming email, detect important messages and route alerts to Telegram. Event-driven async pipeline running in production on home server via Docker.
AI pipeline for removing objects from indoor real-estate images using Stable Diffusion, LoRA fine-tuning and background/foreground segmentation. Includes evaluation with PSNR, SSIM, LPIPS and FID.
Hands-on implementation of qubit state preparation, transmission and measurement using Qiskit. Explores Bloch sphere representation, quantum circuits and noise impact on fidelity.
Ask Me
This assistant is designed to answer quick questions about my background, projects and technical experience, without needing to browse the full portfolio.
It is a self-hosted LLM system, grounded on my CV and project documentation, and built using retrieval-augmented generation (RAG) and prompt orchestration.
It runs entirely on my own hardware, reflecting my interest in understanding and building AI systems end to end, beyond managed cloud services.
Feel free to ask about any project, technology or idea shown on this site.
Contact
Feel free to reach out via email or connect with me on LinkedIn. I'm always open to discussing new projects, ideas, or opportunities.
