aiml.sh — AI Engineering Courses

AI agent engineering courses

I built aiml.sh to teach AI engineering through working examples. Start with model calls and tool calling, then build agents and RAG applications, connect tools with MCP, and evaluate the results with runnable Python code and tests.

The catalog also covers model internals, inference engineering, context and memory, agent harnesses, and MLOps. Courses are free to read, with no sign-up.

Browse courses Start with AI Fundamentals

Courses: 17
Modules: 60
Lessons: 221
Reading hours: ~86

Reading time is estimated. Running examples and completing labs takes additional time.

What you can learn

Courses list their prerequisites and reading duration, with lessons in order. Pick a subject below or browse the courses on aiml.sh.

Foundations and model internals

Start with model requests and structured output. Build a small transformer, then inspect inference, KV caches, scheduling, and capacity.

Agents and workflows

Build agents with tools, state, and multi-step workflows. Connect services, deploy an agent, or stream live audio and screen frames.

MCP and retrieval

Implement MCP servers and clients, secure remote tool access, and build retrieval applications with ingestion, hybrid search, and source references.

Context, memory, and harnesses

Compare control policies, bound model requests and tool actions, and maintain scoped memory with provenance, retention, and deletion rules.

Evaluation and system design

Choose an architecture against quality, latency, cost, and failure requirements. Test RAG and tool calls, then check changes for regressions.

MLOps and releases

Build reproducible datasets and model artifacts, serve a classifier, and exercise release and monitoring controls locally.

Building MLOps

Read, run, and check the result

The lessons connect explanations and diagrams to code you can inspect and run.

Follow a sequence

Start with AI Fundamentals if you are new to model calls and tools. Other courses state the Python, SQL, HTTP, or framework knowledge they expect.

Work through the example

Follow the setup and code in the lesson. Some examples run locally; others need a model API key or cloud service. Check each lesson’s requirements before starting.

Test behavior and limits

Use examples and tests to examine outputs, failure cases, tool boundaries, retrieval quality, and release decisions. Evaluation is part of building the system.

Inside a lesson

A real screen from the AI Fundamentals course: running inference with a local LLM.

aiml.sh lesson explaining local LLM inference, with its course navigation and lesson content

Read “Run inference with a local LLM” on aiml.sh