What you'll learn

Understanding local AI: model, runtime and interface – and why your data stays in-house

Choosing the right hardware and model, even entirely without a GPU

Installing Ollama on Linux and Windows and running it as a service on the network

Measuring and tuning speed: threads, quantization, context and mixture of experts

Building your own model variants with Modelfiles, such as a Linux admin assistant

Using the model in scripts via the command line, the REST API and the OpenAI API

Setting up Open WebUI for the team: users, knowledge base, web search, MCP and HTTPS

AI in admin work: Bash, PowerShell, logs, systemd and SSH hardening reviewed by AI

Using AI agents locally: Cline, Kilo Code, Claude Code and OpenCode with Ollama

Having a local agent write Terraform, Ansible and Docker Compose

Automating with n8n and Python: alerts, documents and whole folders

Running it safely: backups, updates, prompt injection, hallucinations and costs

Requirements

NO experience with AI models required.

Basic knowledge of Linux and the command line is helpful.

A computer with at least 32 GB of memory (or a VM on Proxmox VE) to reproduce the content yourself – no GPU needed.

Description

This course contains the use of artificial intelligence.Welcome to "Local AI with Ollama & Open WebUI: Run Your Own Private AI"!Are you a sysadmin, DevOps engineer, or IT professional who wants to use AI in your daily work without your logs, configs or customer data ending up in the cloud – with no subscription, no cost per request and full control over models and versions? Then this course is exactly right for you!We don't stop at theory. You set up your own AI machine, as a VM on Proxmox VE or on your own computer, install Ollama on Linux and Windows and run an open language model entirely without a GPU. You measure its speed, find the right number of threads and build your own model variants with a Modelfile, such as a Linux admin assistant.Then you talk to your model from the command line, through the REST API and the OpenAI-compatible API – existing ChatGPT code runs locally with one changed line. With Open WebUI, your team gets its own chat interface with users and roles, a knowledge base (RAG) from your own documents, web search with SearXNG, tools via MCP and HTTPS behind nginx.In the main part, you put the local AI to work in everyday admin tasks: writing and checking Bash and PowerShell scripts, analyzing logs and the Windows event log, fixing systemd services and hardening SSH. On top of that come AI agents: Cline and Kilo Code in the editor, Claude Code and OpenCode in the terminal, a local agent that writes Terraform and Ansible – and an honest comparison of local versus cloud.Finally, you automate with n8n and Python – from an alert with AI analysis on your phone to triaging whole folders – and take care of what counts in operations: backups and updates, security and prompt injection, hallucinations and costs.We demonstrate all of this content within live demos – and not through boring presentations. For every lesson, you get the commands to follow along and the scripts and files shown as a download.Our agenda covers topics such as:Understanding local AI: model, runtime and interface – and why your data stays in-houseChoosing the right hardware and model, even entirely without a GPUInstalling Ollama on Linux and Windows and running it as a service on the networkMeasuring and tuning speed: threads, quantization, context and mixture of expertsBuilding your own model variants with Modelfiles, such as a Linux admin assistantUsing the model in scripts via the command line, the REST API and the OpenAI APISetting up Open WebUI for the team: users, knowledge base, web search, MCP and HTTPSAI in admin work: Bash, PowerShell, logs, systemd and SSH hardening reviewed by AIUsing AI agents locally: Cline, Kilo Code, Claude Code and OpenCode with OllamaHaving a local agent write Terraform, Ansible and Docker ComposeAutomating with n8n and Python: alerts, documents and whole foldersRunning it safely: backups, updates, prompt injection, hallucinations and costsWith this course, you gain directly applicable knowledge to run your own AI on your own hardware – private, without a subscription and for your whole team. And all of that in just a few hours.So, take this opportunity and enroll now to make the most of local AI!

System administrators and DevOps engineers who want to use AI in their daily work without handing data to the cloud.,IT professionals and homelab enthusiasts who want to run their own AI with Ollama, Open WebUI, agents and n8n.