Infrastructure & security engineer.
I build, I break, I document.
Popular guides: Docker • Hardened SSH • NGINX • Fail2Ban
Explore: Skills • Experience • Projects
Discover my background and my passion for technology
Expert in cybersecurity, Linux administration and full stack development. I design secure infrastructures and share my knowledge through free tutorials and training courses.
Projects shipped to production, not demos — application security, local AI and automation
Multi-module web vulnerability scanner built in Clean Architecture PHP. 35 active scanners, AI-assisted analysis via the NVIDIA Build API, exports to HackerOne / Bugcrowd / Intigriti.
Self-hosted AI infrastructure: Qwen3-Coder 30B LLM inference on CPU at 15.9 tokens/s, Kokoro text-to-speech, Whisper transcription and Phoenix observability on Proxmox.
Iterative code-generation agent validated end to end at 74s per iteration, with triple fallback: NVIDIA API, local model, then degraded mode.
Tailored solutions to automate and digitize your business
Automate your business workflows with artificial intelligence. LLM integration, autonomous agents, document processing, internal chatbots and smart data pipelines.
Design and development of lightweight, focused SaaS applications. A simple tool that solves one precise problem, with monthly subscription and recurring revenue.
Digital support for schools in Perpignan and the surrounding area. Technical support, network infrastructure, security and IT fleet management.
Building an AI PC, renting a GPU, choosing your tools or setting up your own private assistant: independent, no-nonsense guides from a practitioner who runs this stack every day.
Explore the guidesPrograms, bundles and formats available in Linux cybersecurity and DevOps.
View training coursesExplore my in-depth guides on cybersecurity, system administration and development.
Cybersecurity news, Linux tips and field-tested lessons learned
The practical guide to picking your GGUF quant in 2026: bits per weight, perplexity impact, imatrix, and a VRAM/quality table. A Llama 3.1 8B drops from 32 GB in F32 to 4.9 GB in Q4_K_M.
Build a privacy-first RAG assistant on your own documents: embeddings, vector DB, chunking and a local LLM. With qwen3-embedding hitting 70.58 on multilingual MTEB, local finally rivals commercial APIs.
An honest comparison of local LLM inference engines in 2026: vLLM hits ~793 tok/s under concurrent load versus ~41 for Ollama, yet at a single user the gap drops below 10%. When to use each.
A question or a project to discuss? Send me a message.
For any question or collaboration, send me a message through the contact form.
Contact me
