// real projects

Case studies

Systems we designed and built end to end. Described in plain words — we’re happy to share the technical details on a call.

🖥️
10services working as one

GPU rental marketplace

A service where owners of powerful machines rent them out. The platform connects machines on its own, runs client workloads, meters usage and issues invoices. Under the hood — ten independent services working as one.

  • Go
  • Kafka
  • Docker
  • Postgres
⚙️
1program — the whole server under control

Server supervisor agent

A small program is installed on a GPU server and takes it under control: starts and stops client workloads, watches hardware health and keeps tasks from interfering with each other.

  • Go
  • gRPC
  • Docker
  • Linux
🌐
10+servers configured with one script

Server fleet with smart load distribution

A system of dozens of servers that spreads incoming traffic between them. If one server is overloaded or goes down — traffic automatically shifts to healthy ones. A new server joins the fleet with a single script.

  • HAProxy
  • load balancing
  • web panel
📡
24/7you learn about problems first

Round-the-clock hardware watch

The system regularly checks the health of dozens of devices across servers and sends a Telegram message the moment something needs attention. The owner learns about problems before their customers do.

  • monitoring
  • Telegram
  • web dashboard
🐧
1 commandfrom an empty server to a working one

A ready server with one command

A script turns a clean server into a fully configured system in minutes: installs everything needed, enables protection and outputs ready-to-use connection details. No manual work — and therefore no mistakes.

  • Docker
  • Linux
  • automation
🔁
~1 minfrom failure to a working site

A website that repairs itself

The system watches the website around the clock. If it goes down — within a minute it rents a new server, fills it with AI-generated content and switches visitors over. All without human involvement.

  • AI
  • auto-recovery
  • GitLab
// tools

Technology stack

What we work with every day — from system administration to integrating neural networks.

Linuxnginxsystemdcloud-initPython / asyncioFastAPITypeScriptBashOllamaGemma / LlamaOpenClawTelegram Bot APIGitLab CI/CDCloudflare APITimeWeb CloudSQLite / SQLAlchemyJWT / bcryptVue / ViteLet's Encryptpytest

Want a system like this?

Tell us what your project needs — we’ll propose a solution and an estimate within a day.