# Local AI Master

> Courses and guides for running AI on hardware you own: 25 hands-on courses (561 chapters) on local LLMs, RAG, agents, fine-tuning, speech and deployment with Ollama, vLLM and open models, plus an in-browser Python Lab. Guides are free to read with no account. The first chapter of every course is free with a free account (no card). Full access is Pro at $8.99/month (cancel any time) or Lifetime at $149 paid once, which includes every future course. All sales are final.

## Courses (public descriptions, syllabus and pricing — no login needed)
- [What is AI](https://localaimaster.com/courses/browse/what-is-ai): A foundations course that explains what artificial intelligence is at a mechanical level, from pattern recognition and tokenization through attention, training and fine-tuning, for someone starting with no machine learni (142 chapters, about 45 hours)
- [AI: Beginning to Advanced](https://localaimaster.com/courses/browse/ai-beginning-to-advanced): A deep learning course that starts by building a neuron and an autograd engine by hand, then works forward through the architecture families, GPU behavior, and the systems work required to keep a trained model running. (39 chapters, about 30 hours)
- [Natural Language Processing](https://localaimaster.com/courses/browse/nlp): A working course in natural language processing that treats text as an engineering problem rather than a prompting exercise, covering tokenization, embeddings, sequence labeling, retrieval, generation and the evaluation  (14 chapters, about 14 hours)
- [Computer Vision](https://localaimaster.com/courses/browse/computer-vision): A computer vision course that starts by building a convolutional network from first principles and ends with detection, segmentation, video, 3D and vision-language systems running on constrained hardware, with a heavy em (13 chapters, about 13 hours)
- [RAG Systems](https://localaimaster.com/courses/browse/rag-system): A hands-on course on retrieval-augmented generation for developers who need a language model to answer accurately from a private corpus: how to ingest and chunk documents, choose and operate embeddings and a vector store (11 chapters, about 12 hours)
- [Agentic AI](https://localaimaster.com/courses/browse/agentic-ai): A design-and-build course on autonomous AI agents: how the agent loop actually works, how to specify tools a model can use without supervision, which planning and reasoning patterns hold up, how memory and state are mana (10 chapters, about 11 hours)
- [Multimodal AI Systems](https://localaimaster.com/courses/browse/multimodal-ai): An architecture-first course on multimodal systems that works through the actual designs used to connect perception to language — the projector, the Q-Former, cross-attention resamplers, natively multimodal training, and (8 chapters, about 9 hours)
- [MLOps](https://localaimaster.com/courses/browse/mlops): A practical course on the engineering discipline that keeps machine learning systems working after launch: reproducible pipelines, model registries, serving architecture, drift monitoring, safe rollout, automated recover (12 chapters, about 14 hours)
- [Test-Time Compute Scaling](https://localaimaster.com/courses/browse/test-time-compute): A focused course on the second scaling axis — improving answer quality by spending more computation when the model runs, rather than by training a larger model. It covers sequential reasoning, parallel sampling, verifica (3 chapters, about 4 hours)
- [Reinforcement Learning](https://localaimaster.com/courses/browse/reinforcement-learning): A sequential-decision-making course that begins with bandits and Markov decision processes, works through value methods, policy gradients and model-based control, and ends with RLHF and the operational realities of runni (12 chapters, about 14 hours)
- [MCP Servers and Tool Ecosystems](https://localaimaster.com/courses/browse/mcp-servers): A practical course on the Model Context Protocol for engineers who need to expose real systems to AI clients: how the protocol is structured, how tools and resources differ, which transport to run, how authorization and  (12 chapters, about 10 hours)
- [Voice AI and Realtime Agents](https://localaimaster.com/courses/browse/voice-ai): An engineering course in building conversational voice agents that answer in real time, covering the audio pipeline, turn-taking, tool use, telephony integration and the evaluation and operations work that keeps one runn (12 chapters, about 10 hours)
- [AI Engineering: From Prompt to Production](https://localaimaster.com/courses/browse/ai-engineering): A course on building applications around a language model rather than around a demo — output contracts, retrieval, evaluation, guardrails, cost and latency control, and the operational work that keeps an LLM feature work (12 chapters, about 11 hours)
- [AI Security, Guardrails, and Red Teaming](https://localaimaster.com/courses/browse/ai-security): A course on securing systems built with language models, covering threat modeling, prompt injection, insecure output handling, agent permissions, guardrail design, adversarial testing, monitoring and the governance contr (12 chapters, about 11 hours)
- [Human-AI Collaboration](https://localaimaster.com/courses/browse/human-ai-collaboration): A course on designing the working relationship between people and AI systems — which decisions get delegated, what review actually catches errors, how trust should be calibrated to evidence, and how to roll any of it out (12 chapters, about 9 hours)
- [Local AI Deployment: From Laptop to Production](https://localaimaster.com/courses/browse/local-ai-deployment): A hands-on course on operating large language models on hardware you own, covering quantization, KV cache sizing, runtime selection, multi-GPU and edge deployment, and putting an OpenAI-compatible API in front of it all. (16 chapters, about 18 hours)
- [Dataset Engineering: Build the Data That Makes Models Great](https://localaimaster.com/courses/browse/dataset-engineering): A course on the engineering discipline that produces training corpora — sourcing, filtering, deduplicating, generating, labelling, evaluating and versioning the data that a model is trained on, with production code rathe (16 chapters, about 16 hours)
- [AI for Coding, Code Agents, and Engineering Workflows](https://localaimaster.com/courses/browse/ai-for-coding): A hands-on course on building and operating AI systems that work inside a real repository — retrieving code context, editing files through diffs, running tests, and asking for human approval before anything risky — aimed (12 chapters, about 56 hours)
- [Fine-Tuning, Distillation, and Model Adaptation](https://localaimaster.com/courses/browse/fine-tuning-distillation): A course on adapting language models to a specific job — choosing between prompting, distillation, supervised fine-tuning and preference optimization, building the dataset that makes any of them work, and proving the res (12 chapters, about 50 hours)
- [Local Image Generation: Stable Diffusion, FLUX & ComfyUI](https://localaimaster.com/courses/browse/image-generation): A hands-on course in running diffusion image models on your own GPU, using ComfyUI as the workbench and SDXL and FLUX as the engines, for people who want repeatable results rather than lucky ones. (12 chapters, about 12 hours)
- [Digital Marketing: Zero to Pro](https://localaimaster.com/courses/browse/digital-marketing): A full-stack digital marketing course covering positioning and brand, websites and landing pages, SEO, content, social, email, paid advertising, analytics and attribution, and where AI genuinely helps a marketer. (115 chapters, about 40 hours)
- [AI Hardware: Build, Buy & Size Your Local AI Machine](https://localaimaster.com/courses/browse/ai-hardware): A buying and sizing course for people about to spend real money on a machine to run AI locally: why memory capacity gates what runs at all, why memory bandwidth sets how fast it runs, how the accelerator platforms genuin (12 chapters, about 9 hours)
- [Ollama Mastery: From Install to Daily Driver](https://localaimaster.com/courses/browse/ollama-mastery): An operations course for the tool most people actually run local models with: how Ollama loads, splits and unloads models, what a tag commits you to, why the GPU is often doing none of the work, and how to expose the ser (15 chapters, about 10 hours)
- [Local Speech Studio: TTS, Voice Cloning and Transcription](https://localaimaster.com/courses/browse/local-speech-studio): A production-oriented course in running speech synthesis, voice cloning and transcription on your own hardware, covering the engines, the pipelines that turn them into finished audio, and the license terms that decide wh (13 chapters, about 9 hours)
- [Pick the Right Model for Your Machine](https://localaimaster.com/courses/browse/local-model-picker): A decision course for people who keep downloading models and cannot tell which one is actually better: picks organised by capacity tier and by job to start from, then the reasoning underneath them — the memory arithmetic (14 chapters, about 9 hours)

## Pricing and access
- [Pricing](https://localaimaster.com/pricing): Pro $8.99/month or $79/year; Lifetime $149 once. Both unlock all 25 courses, the Python Lab, quizzes and certificates.
- [Create a free account](https://localaimaster.com/auth/register): reads chapter 1 of every course. Course chapters beyond the first are for paying members and are not public.

## Key guides & tools (free, no account)
- [Best Ollama models](https://localaimaster.com/blog/best-ollama-models)
- [Ollama system requirements](https://localaimaster.com/blog/ollama-system-requirements)
- [Hardware guide hub](https://localaimaster.com/hardware)
- [Best LLM for 8 GB VRAM](https://localaimaster.com/vram/best-llm-8gb-vram)
- [Best LLM for 16 GB VRAM](https://localaimaster.com/vram/best-llm-16gb-vram)
- [Best LLM for 24 GB VRAM](https://localaimaster.com/vram/best-llm-24gb-vram)
- [Best coding LLM for 16 GB VRAM](https://localaimaster.com/vram/best-coding-llm-16gb-vram)
- [Best local TTS models](https://localaimaster.com/blog/best-local-tts-models)
- [Run FLUX.1 locally](https://localaimaster.com/blog/flux-local-image-generation)
- [Build local AI agents](https://localaimaster.com/blog/ai-agents-local-guide)
- [Model directory](https://localaimaster.com/models)
- [VRAM calculator](https://localaimaster.com/tools/vram-calculator)
- [Model recommender](https://localaimaster.com/tools/model-recommender)
- [Blog](https://localaimaster.com/blog)

## About
- [About](https://localaimaster.com/about) · [Contact](https://localaimaster.com/contact) · [Terms](https://localaimaster.com/terms) · [Privacy](https://localaimaster.com/privacy)