Overview
I set up and maintain on-premise infrastructure for organizations that need full control over their data, cannot move to the cloud due to regulatory requirements, or want predictable hosting costs. From bare-metal Linux servers to Docker Swarm clusters running local AI models, I handle the full stack - with monitoring, backup, and disaster recovery built in.
What I Can Do For You
Server Setup and Administration
- Linux server provisioning: Ubuntu, Debian, Rocky Linux - secure baseline, firewall, fail2ban, auditing
- Docker orchestration: Compose for small deployments, Docker Swarm or Nomad for clusters
- Reverse proxy: Nginx, Caddy, or Traefik - TLS termination, virtual hosting, rate limiting
- Backup and disaster recovery: Automated off-site backups, snapshot rotation, restore drills
Local AI Deployment
- LM Studio / Ollama: Deploy open-source LLMs (Llama, Mistral, Qwen) on local hardware
- GPU acceleration: NVIDIA CUDA setup, container runtime configuration, model quantization
- API gateways: Expose local models via OpenAI-compatible APIs for your applications to consume
- Air-gapped environments: Fully offline AI deployment with no external dependencies
Infrastructure Management
- Monitoring: Prometheus + Grafana stack, uptime monitoring, alerting
- Logging: Centralized logging with Loki or ELK stack
- VPN: WireGuard or Tailscale for secure remote access
- CI/CD self-hosted: GitLab Runner or Jenkins on your own hardware
Why On-Premise?
- Data sovereignty: Sensitive data never leaves your network
- Predictable costs: Fixed hardware costs - no cloud egress surprises
- Latency: Sub-millisecond access for local AI inference
- Compliance: Meet GDPR, HIPAA, SOC 2 requirements without cloud dependency
Technologies
OS: Linux (Ubuntu, Debian, Rocky Linux, Arch) Containers: Docker, Docker Compose, Docker Swarm AI: LM Studio, Ollama, local LLMs Web: Nginx, Caddy, Traefik Monitoring: Prometheus, Grafana, Uptime Kuma Networking: WireGuard, Tailscale, iptables, nftables Backup: Borg, Restic, rsync