One course for AI engineering, agents, and prompting.

What AI engineering is, the roles it covers, and the baseline skills you need before diving in.

How large language models actually work: transformers, tokens, training, inference, and their limits.

The major model providers, open-weight families, local runtimes, and how to choose (and pay for) a model.

Writing effective prompts: clarity, context, examples, roles, output control, and the core prompting techniques.

Beyond single prompts: ensembles, self-consistency, automatic prompt engineering, and managing the context window deliberately.

Calling models in real code: provider APIs and SDKs, streaming, sampling parameters, and cost-saving techniques.

Turning text into vectors: embedding models, similarity search, and the vector stores that power semantic retrieval.

Grounding model answers in your own data: chunking, retrieval, generation, and when RAG beats fine-tuning.

From single prompts to autonomous loops: what agents are, where they shine, and the core agent architectures.

Giving models the ability to act: tool definitions, provider function-calling APIs, and the common tool categories.

How agents remember: short vs long-term memory, episodic vs semantic stores, and strategies for summarizing and forgetting.

Building agents with (or without) a framework: LangChain, LlamaIndex, LangGraph, CrewAI, and the provider agent SDKs.

The open standard for connecting models to tools and data: MCP architecture, and building your own servers and clients.

Beyond text: vision, image generation, video, and speech — and the APIs and frameworks that combine them.

Knowing whether your AI system works: metrics, testing strategies, eval frameworks, and production tracing.

Shipping AI responsibly: prompt injection, red teaming, sandboxing, privacy, moderation, bias, and ethics.

The AI-powered development tools reshaping how software gets built — and how to work with them effectively.