Local AI Revolution: Ollama and OpenClaw

Overview

Take full control of your AI by building a complete self-hosted stack from scratch using Ollama and OpenClaw. Learn how to serve open models, accelerate inference with rented GPUs, and deploy autonomous AI agents that work on your terms.Deploy Ollama on a Linux serverRun open-weight LLMs privatelyBuild a ChatGPT-style interface with Open WebUIMove AI workloads from CPU to GPURent and configure GPU infrastructureInstall and configure OpenClawBuild autonomous agentic AI workflowsConnect AI agents to Telegram

Syllabus

•   Introduction
Introduction to the Local AI Revolution
Exercise: Meet Your Classmates and Instructor
Course Resources
•   Private AI: Ollama + Open Source Models From Scratch
DigitalOcean Sign Up and SSH Setup
Create the Droplet and Set Up Your User
Install and Configure Ollama
Where to Find Ollama Models: Ollama Library, Hugging Face, GGUF, and Model Tags
Pull Models and Run Your First Inference
Parameters, Quantization, and VRAM: Why Models Are the Size They Are
Adjusting Model Behaviour in the Interactive Shell
Saving and Reusing Model Configurations
Customizing Models with Modelfile
ChatGPT-Like Interface for Your Private AI: Open WebUI Setup
Access Your AI Securely From Anywhere: SSH Tunneling Explained
LM Studio: A Desktop Alternative to Ollama
•   GPU Inference on Vast.ai + OpenClaw: 20× Faster, Same Stack
Introduction: From CPU to GPU
Sign Up on Vast.ai and Add Your SSH Key
Deploy a GPU Instance and Connect
Install and Configure Ollama
Pull Models and Manage the Instance
CPU vs GPU - Live Benchmark
What Is Agentic AI? The Shift From Chatbots to Autonomous Agents
Install and Configure OpenClaw
Session Startup and Terminal Multiplexing
Connect Telegram to OpenClaw
OpenClaw Config: Telegram Hardening
•   Where To Go From Here?
Review This Byte!
Become An Alumni

---

Taught by

Andrei Dumitrescu

---

Subjects

Artificial Intelligence