COURSE 10Intermediate9 IN-DEPTH LESSONS

Agents and Tool-Using Systems

Autonomous systems that loop: tool calling, function interfaces, short and long-term memory, multi-step planning, and human-in-the-loop safety gates.

Start Lesson 1: AI agents →← All 29 Courses

Course Syllabus & Units

Unit 1

Loops

1 lessons
1.1L1

AI agents

An **AI agent** (in today’s product sense) is a system that **loops**: observe → decide → act with tools → check → repeat toward a goal. A plain chatbot answers once. An agent may browse, call APIs, write files, or ask you for confirmati...

Read lesson →
Unit 2

Tools

1 lessons
2.1L1

Tool use and function calling

**Tool use** lets a model call external functions—search, calculators, calendars, databases—instead of only emitting prose. **Function calling** is a common API pattern: the model returns a structured “please call this tool with these ar...

Read lesson →
Unit 3

Memory

1 lessons
3.1L1

Memory for agents

Agents need memory beyond a single message: the current chat (short-term), and sometimes durable stores (preferences, project files, retrieved docs).

Read lesson →
Unit 4

Evaluation

1 lessons
4.1L1

Evaluation of agents

**Evaluating agents** means checking whether a system that plans, uses tools, and takes multi-step actions actually finishes the job—safely, reliably, and at acceptable cost.

Read lesson →
Unit 5

Planning

1 lessons
5.1L1

Planning and multi-step reasoning

**Planning** breaks a goal into steps. **Multi-step reasoning** is the model (or system) working through those steps—sometimes with tools—before a final answer.

Read lesson →
Unit 6

Multi

1 lessons
6.1L1

Multi-agent patterns

Designs where multiple agent roles collaborate—researcher, critic, executor—via messages or shared state.

Read lesson →
Unit 7

Permissions

1 lessons
7.1L1

Permissions and approvals

Least-privilege tool access and **human approval** gates for risky agent actions.

Read lesson →
Unit 8

Orchestration

1 lessons
8.1L1

Workflow orchestration

Running multi-step AI work as an explicit **workflow**: graphs of steps with retries, timeouts, state, and human gates—not only a free-form chat loop.

Read lesson →
Unit 9

Reliability

1 lessons
9.1L1

Agent timeouts and budgets

Hard limits on agent runs: max steps, max tool calls, max tokens, wall-clock timeouts, and spend caps.

Read lesson →