LLM Fundamentals
The key concepts behind modern AI explained for beginners: AI, machine learning, and neural networks, large and small language models, vision-language and multimodal models, frontier models and AGI, parameters and weights, tokens and tokenizers, the context window, and what goes into a model's context.
Start with AI, Machine Learning, and Neural Networks →Kinds of Models
What AI, machine learning, and neural networks are, and the kinds of models built on them: large and small language models, vision-language models, multimodal assistants, frontier models, and AGI.
How Models Work
What a model is made of and what it reads: parameters and weights, tokens, tokenizers, the context window, and the context sent with every request.
Training Models
How models are built and changed: training and inference, pre-training and post-training, fine-tuning, LoRA, synthetic data, RLHF, distillation, and data poisoning.
Running Models
What affects speed, cost, and quality when a model runs: reasoning models, caching, quantization, and model routing.
Prompts and APIs
How to talk to a model: prompts, prompt engineering, context engineering, structured outputs, and APIs for AI models.
Agents
How a model becomes an agent: models versus agents, tools, the agent loop, and the harness.
Knowledge and Memory
How models get facts right and keep track of long work: hallucinations, grounding, RAG, vector databases and embeddings, memory, compaction, and context rot.
Agent Tooling
What agents are connected to and how people work with them: MCP, agent skills, computer use, sandboxes, agentic search, human in the loop, agentic engineering, and vibe coding.
Quality and Safety
How models are measured and attacked: benchmarks, evals, prompt injection, data exfiltration, and open-weights and open-source models.
Interview Practice
Interview questions on every topic of the course: concept questions with short answers and follow-ups, and scenario, design, and spot-the-mistake questions.