# 0x0800 SRL | Denis Leclercq

> On-premise infrastructure, cloud, and local AI | Belgium and France

## About

0x0800 SRL (Denis Leclercq, Liège, Belgium, BCE: BE1021845609) builds and operates
infrastructure across three fields treated as one discipline: on-premise systems, cloud and
hybrid platforms, and local AI inference. Twenty years of infrastructure work back the
delivery, from network support to technical direction. The company publishes its own
measurements: benchmarks are run on in-house hardware and released with their method,
commands, and caveats.

## What we do

- On-premise infrastructure: virtualised estates (Proxmox, VMware), networking and BGP routing,
  high availability, backup, supervision, redundant databases and front ends, operational
  documentation.
- Cloud and hybrid: scalable AWS and Azure architectures, migration of on-premise environments
  to the cloud, infrastructure as code (Terraform, Ansible), CI/CD pipelines with DevSecOps
  integration, continuous drift detection and compliance evidence.
- Local AI: vLLM inference servers and quantised models, memory/thermal/context sizing, agents
  connected to internal sources, Hygur local edition deployment, team training and handover.

Pricing is not published. Work is quoted per engagement after an on-site scoping day.

## Method

1. Scoping: one day on site to review the existing setup, the intended use, and the data and
   compliance constraints. Output: a roadmap and a quote.
2. Build: installation, migration, or deployment, depending on the engagement. A narrow scope
   first, production second.
3. Autonomy: documentation, team training, supervision. No dependency created.

## Articles

Measurements published from in-house hardware, reproducible command by command.

- DeepSeek-V4-Flash-0731 on two DGX Spark: the complete recipe (4 August 2026) —
  https://0x0800.com/article-dsv4-2x-spark-recette.html
  End-to-end tensor-parallel setup on two GB10 boxes: RoCE networking, build, checkpoint transfer,
  launch. Then the measurements: code quality judged by execution (96/100 over 100 draws),
  throughput by workload, session capacity, thermals and RDMA interconnect latency, plus the
  traps that each cost an hour.
- DSv4 on 2x DGX Spark: assembly, measurements, tuning (4 August 2026) —
  https://0x0800.com/article-dsv4-2x-spark-rapport.html
  The reproducible companion report: both software paths end to end, the +31 s first-token
  latency cliff introduced by the b12x MoE backend, the one-knob sweep, and the single-parameter
  law that caps single-stream throughput. Twelve parameters discarded out of sixteen.
- Serving Laguna S 2.1 NVFP4 on a single DGX Spark (23 July 2026) —
  https://0x0800.com/article-laguna-s21-single-spark.html
  A 117.6B-parameter MoE on one unified-memory box: memory hygiene, native 256K context,
  DFlash speculative decoding, throughput benchmarks.
- Running vLLM properly on a DGX Spark, GB10 (20 July 2026) —
  https://0x0800.com/article-dgx-spark-vllm-nvfp4.html
  The three bugs that block the build and first server on SM121, and how to fix them.

## Products

- Hygur (https://hygur.ai): Local-first personal AI. Data stays on the machine, the model comes
  to it. Deployed on-premise in companies, in local edition and sovereign cloud.
- CompliMetric (https://complimetric.com): Continuous cloud compliance. Scans cloud and IaC,
  detects drift, produces NIS2 / ISO 27001 evidence. SaaS + GitHub Action + MCP server.

Both are edited and operated in-house.

## Contact

- Email: contact@0x0800.com
- Website: https://0x0800.com
- LinkedIn: https://www.linkedin.com/in/denis-leclercq
- GitHub: https://github.com/Mirrdhyn
- Location: Liège, Belgium (serves Belgium, France, EU)

## Full Profile

For detailed information, see: https://0x0800.com/llms-full.txt