🖥 Technical

Hosting

Hosted by OVH — VPS-1 (4 vCPU, 8 GB RAM, 75 GB SSD NVMe) in Roubaix, France (EU-WEST-RBX).
Static site + Flask API on Docker: nginx reverse proxy, gunicorn WSGI server, TLS via Let's Encrypt (Certbot).

Current snapshot

  • Disk: —
  • RAM: —
  • CPU: —
  • Load (1 / 5 / 15 min): —
  • Uptime: —

Docker containers (this site)

—

Capacity timeline

disk   RAM   CPU

Usage renormalized to contracted capacity (75 GB disk, 8 GB RAM, 4 vCPU). The gap to 100% is headroom. The VPS records RAM, CPU, and disk every 15 minutes; the snapshot and timeline are read live from that log.

Anonymized connections by country

Last daily automatic update on —. Counts are per unique IP: — visitors.

Graph of outbound links from this site (centre) and from the sites it references. Double-click a node to open the corresponding page; drag nodes to move.

About development

This website, like many of my most recent projects, is partly built using various generative models. I believe that using those powerful tools today is becoming a standard and somehow a necessity for efficiency in many computer science–related jobs, even in research.

I started coding before the global usage of AI. I believe the new tools—Claude Code by Anthropic, Cursor IDE, and all the various fast-evolving models, each with their specificities—should be used as powerful TOOLS. And like every tool, they need thorough usage and practice to be efficient. I always try as much as I can to keep a critical look at outputs, understand main architecture choices, key dependencies, and unplug from time to time to convince myself I still know how to code an efficient multithreaded C++ local search.