$ lcpp status
profile performance
model Qwen3-8B-Q4_K_M
tok/s 59.5
uptime 6d 14h
● llama-server.service active (running)

Stefan Syrett

Platform engineer. Self-hosted everything. If it's worth doing, it's worth doing with a health check.

tool · open source

llama-launcher

Profile-driven CLI for llama.cpp running as a systemd --user service. Switch inference profiles, autotune context from available VRAM, game-aware watch daemon that hands the GPU back when you launch a game. No sudo needed anywhere.

25+
commands
0
sudo required
CUDA + ROCm
backends
PythonTypersystemdllama.cpppipx
lcpp --help output listing the available commands

GitHub mirror ↗

open project →

product · private beta

parsed.gg

Every CS2 match, broken down automatically. 2D round replays, stats that mean something, and AI coaching that talks tactics — straight from your Steam matchmaking history.

Steam
auto-sync
2D
round replay
AI
round coaching
FastAPIReactPostgresRedisDocker
parsed.gg stats page showing career K/D and map breakdown

Private beta — invite-only for now.

Join waitlist ↗ open project →

data pipeline · open source

UK house prices

Full historical ingest of HM Land Registry Price Paid Data, enriched with EPC energy certificates for floor area and £/m². Postcode geocoding, Grafana dashboards, and a conversational analytics agent a local LLM drives over MCP — kept current with automated monthly updates.

28M+
transactions
1995–present
coverage
monthly
auto-update
PostgreSQLFastAPIGrafanallama.cppQwen3-8BMCP
Grafana property price dashboard — inside the M25Grafana property price dashboard — outside the M25 (Kent)

GitHub ↗

open project →

day job

  • Azure · Terraform · AKS
  • Python · FastAPI · Postgres
  • DevSecOps
  • Regulated env (NHS/CQC-adj.)

homelab / side

  • Docker · systemd
  • llama.cpp · n8n · Grafana
  • CachyOS · Debian · btrfs
  • Self-hosted LLM (RTX 3060)

Platform engineer at a regulated UK public body. Open to senior infrastructure, DevOps, and platform roles.